gcc/libgomp/allocator.c

1472 lines
42 KiB
C
Raw Normal View History

2024-01-03 19:19:35 +08:00
/* Copyright (C) 2020-2024 Free Software Foundation, Inc.
Contributed by Jakub Jelinek <jakub@redhat.com>.
This file is part of the GNU Offloading and Multi Processing Library
(libgomp).
Libgomp is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 3, or (at your option)
any later version.
Libgomp is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
FOR A PARTICULAR PURPOSE. See the GNU General Public License for
more details.
Under Section 7 of GPL version 3, you are granted additional
permissions described in the GCC Runtime Library Exception, version
3.1, as published by the Free Software Foundation.
You should have received a copy of the GNU General Public License and
a copy of the GCC Runtime Library Exception along with this program;
see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
<http://www.gnu.org/licenses/>. */
/* This file contains wrappers for the system allocation routines. Most
places in the OpenMP API do not make any provision for failure, so in
general we cannot allow memory allocation to fail. */
#define _GNU_SOURCE
#include "libgomp.h"
#include <stdlib.h>
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
#include <string.h>
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
#include <dlfcn.h>
#endif
OpenMP/Fortran: Implement omp allocators/allocate for ptr/allocatables This commit adds -fopenmp-allocators which enables support for 'omp allocators' and 'omp allocate' that are associated with a Fortran allocate-stmt. If such a construct is encountered, an error is shown, unless the -fopenmp-allocators flag is present. With -fopenmp -fopenmp-allocators, those constructs get turned into GOMP_alloc allocations, while -fopenmp-allocators (also without -fopenmp) ensures deallocation and reallocation (via intrinsic assignments) are properly directed to GOMP_free/omp_realloc - while normal Fortran allocations are processed by free/realloc. In order to distinguish a 'malloc'ed from a 'GOMP_alloc'ed memory, the version field of the Fortran array discriptor is (mis)used: 0 indicates the normal Fortran allocation while 1 denotes GOMP_alloc. For scalars, there is record keeping in libgomp: GOMP_add_alloc(ptr) will add the pointer address to a splay_tree while GOMP_is_alloc(ptr) will return true it was previously added but also removes it from the list. Besides Fortran FE work, BUILT_IN_GOMP_REALLOC is no part of omp-builtins.def and libgomp gains the mentioned two new function. gcc/ChangeLog: * builtin-types.def (BT_FN_PTR_PTR_SIZE_PTRMODE_PTRMODE): New. * omp-builtins.def (BUILT_IN_GOMP_REALLOC): New. * builtins.cc (builtin_fnspec): Handle it. * gimple-ssa-warn-access.cc (fndecl_alloc_p, matching_alloc_calls_p): Likewise. * gimple.cc (nonfreeing_call_p): Likewise. * predict.cc (expr_expected_value_1): Likewise. * tree-ssa-ccp.cc (evaluate_stmt): Likewise. * tree.cc (fndecl_dealloc_argno): Likewise. gcc/fortran/ChangeLog: * dump-parse-tree.cc (show_omp_node): Handle EXEC_OMP_ALLOCATE and EXEC_OMP_ALLOCATORS. * f95-lang.cc (ATTR_ALLOC_WARN_UNUSED_RESULT_SIZE_2_NOTHROW_LIST): Add 'ECF_LEAF | ECF_MALLOC' to existing 'ECF_NOTHROW'. (ATTR_ALLOC_WARN_UNUSED_RESULT_SIZE_2_NOTHROW_LEAF_LIST): Define. * gfortran.h (gfc_omp_clauses): Add contained_in_target_construct. * invoke.texi (-fopenacc, -fopenmp): Update based on C version. (-fopenmp-simd): New, based on C version. (-fopenmp-allocators): New. * lang.opt (fopenmp-allocators): Add. * openmp.cc (resolve_omp_clauses): For allocators/allocate directive, add target and no dynamic_allocators diagnostic and more invalid diagnostic. * parse.cc (decode_omp_directive): Set contains_teams_construct. * trans-array.h (gfc_array_allocate): Update prototype. (gfc_conv_descriptor_version): New prototype. * trans-decl.cc (gfc_init_default_dt): Fix comment. * trans-array.cc (gfc_conv_descriptor_version): New. (gfc_array_allocate): Support GOMP_alloc allocation. (gfc_alloc_allocatable_for_assignment, structure_alloc_comps): Handle GOMP_free/omp_realloc as needed. * trans-expr.cc (gfc_conv_procedure_call): Likewise. (alloc_scalar_allocatable_for_assignment): Likewise. * trans-intrinsic.cc (conv_intrinsic_move_alloc): Likewise. * trans-openmp.cc (gfc_trans_omp_allocators, gfc_trans_omp_directive): Handle allocators/allocate directive. (gfc_omp_call_add_alloc, gfc_omp_call_is_alloc): New. * trans-stmt.h (gfc_trans_allocate): Update prototype. * trans-stmt.cc (gfc_trans_allocate): Support GOMP_alloc. * trans-types.cc (gfc_get_dtype_rank_type): Set version field. * trans.cc (gfc_allocate_using_malloc, gfc_allocate_allocatable): Update to handle GOMP_alloc. (gfc_deallocate_with_status, gfc_deallocate_scalar_with_status): Handle GOMP_free. (trans_code): Update call. * trans.h (gfc_allocate_allocatable, gfc_allocate_using_malloc): Update prototype. (gfc_omp_call_add_alloc, gfc_omp_call_is_alloc): New prototype. * types.def (BT_FN_PTR_PTR_SIZE_PTRMODE_PTRMODE): New. libgomp/ChangeLog: * allocator.c (struct fort_alloc_splay_tree_key_s, fort_alloc_splay_compare, GOMP_add_alloc, GOMP_is_alloc): New. * libgomp.h: Define splay_tree_static for 'reverse' splay tree. * libgomp.map (GOMP_5.1.2): New; add GOMP_add_alloc and GOMP_is_alloc; move GOMP_target_map_indirect_ptr from ... (GOMP_5.1.1): ... here. * libgomp.texi (Impl. Status, Memory management): Update for allocators/allocate directives. * splay-tree.c: Handle splay_tree_static define to declare all functions as static. (splay_tree_lookup_node): New. * splay-tree.h: Handle splay_tree_decl_only define. (splay_tree_lookup_node): New prototype. * target.c: Define splay_tree_static for 'reverse'. * testsuite/libgomp.fortran/allocators-1.f90: New test. * testsuite/libgomp.fortran/allocators-2.f90: New test. * testsuite/libgomp.fortran/allocators-3.f90: New test. * testsuite/libgomp.fortran/allocators-4.f90: New test. * testsuite/libgomp.fortran/allocators-5.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-14.f90: Add coarray and not-listed tests. * gfortran.dg/gomp/allocate-5.f90: Remove sorry dg-message. * gfortran.dg/bind_c_array_params_2.f90: Update expected dump for dtype '.version=0'. * gfortran.dg/gomp/allocate-16.f90: New test. * gfortran.dg/gomp/allocators-3.f90: New test. * gfortran.dg/gomp/allocators-4.f90: New test.
2023-12-08 22:18:25 +08:00
/* Keeping track whether a Fortran scalar allocatable/pointer has been
allocated via 'omp allocators'/'omp allocate'. */
struct fort_alloc_splay_tree_key_s {
void *ptr;
};
typedef struct fort_alloc_splay_tree_node_s *fort_alloc_splay_tree_node;
typedef struct fort_alloc_splay_tree_s *fort_alloc_splay_tree;
typedef struct fort_alloc_splay_tree_key_s *fort_alloc_splay_tree_key;
static inline int
fort_alloc_splay_compare (fort_alloc_splay_tree_key x, fort_alloc_splay_tree_key y)
{
if (x->ptr < y->ptr)
return -1;
if (x->ptr > y->ptr)
return 1;
return 0;
}
#define splay_tree_prefix fort_alloc
#define splay_tree_static
#include "splay-tree.h"
#define splay_tree_prefix fort_alloc
#define splay_tree_static
#define splay_tree_c
#include "splay-tree.h"
static struct fort_alloc_splay_tree_s fort_alloc_scalars;
/* Add pointer as being alloced by GOMP_alloc. */
void
GOMP_add_alloc (void *ptr)
{
if (ptr == NULL)
return;
fort_alloc_splay_tree_node item;
item = gomp_malloc (sizeof (struct splay_tree_node_s));
item->key.ptr = ptr;
item->left = NULL;
item->right = NULL;
fort_alloc_splay_tree_insert (&fort_alloc_scalars, item);
}
/* Remove pointer, either called by FREE or by REALLOC,
either of them can change the allocation status. */
bool
GOMP_is_alloc (void *ptr)
{
struct fort_alloc_splay_tree_key_s needle;
fort_alloc_splay_tree_node n;
needle.ptr = ptr;
n = fort_alloc_splay_tree_lookup_node (&fort_alloc_scalars, &needle);
if (n)
{
fort_alloc_splay_tree_remove (&fort_alloc_scalars, &n->key);
free (n);
}
return n != NULL;
}
#define omp_max_predefined_alloc omp_thread_mem_alloc
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
#define ompx_gnu_min_predefined_alloc ompx_gnu_pinned_mem_alloc
#define ompx_gnu_max_predefined_alloc ompx_gnu_pinned_mem_alloc
OpenMP: Allocate directive for static vars, clean up For the 'allocate' directive, remove the sorry for static variables and just keep using normal memory, but honor the requested alignment and set a DECL_ATTRIBUTE in case a target may want to make use of this later on. The documentation is updated accordingly. The C diagnostic to check for predefined allocators (req. for static vars) failed to accept GCC's ompx_gnu_... allocator, now fixed. (Fortran was already okay; but both now use new common #defined value for checking.) And while Fortran common block variables are still rejected, the check has been improved as before the sorry diagnostic did not work for common blocks in modules. Finally, for 'allocate' clause on the target/task/taskloop directives, there is now a warning for omp_thread_mem_alloc (i.e. predefined allocator with access = thread), which is undefined behavior according to the OpenMP specification. And, last, testing showed that var decl + static_assert sets TREE_USED but does not produce a statement list in C, which did run into an assert in gimplify. This special case is now also handled. gcc/c/ChangeLog: * c-parser.cc (c_parser_omp_allocate): Set alignment for alignof; accept static variables and fix predef allocator check. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Use gomp-constants.h consts. * trans-common.cc (translate_common): Reject OpenMP allocate directives. * trans-decl.cc (gfc_finish_var_decl): Handle allocate directive for static variables. (gfc_trans_deferred_vars): Update for the latter. gcc/ChangeLog: * gimplify.cc (gimplify_bind_expr): Fix corner case for OpenMP allocate directive. (gimplify_scan_omp_clauses): Warn if omp_thread_mem_alloc is used as allocator with the target/task/taskloop directive. include/ChangeLog: * gomp-constants.h (GOMP_OMP_PREDEF_ALLOC_MAX, GOMP_OMPX_PREDEF_ALLOC_MIN, GOMP_OMPX_PREDEF_ALLOC_MAX, GOMP_OMP_PREDEF_ALLOC_THREADS): New defines. libgomp/ChangeLog: * allocator.c: Add static asserts for news GOMP_OMP{,X}_PREDEF_ALLOC_{MIN,MAX} range values. * libgomp.texi (OpenMP Impl. Status): Allocate directive for static vars is now supported. Refer to PR for allocate clause. (Memory allocation): Update for static vars; minor word tweaking. gcc/testsuite/ChangeLog: * c-c++-common/gomp/allocate-9.c: Update for removed sorry. * gfortran.dg/gomp/allocate-15.f90: Likewise. * gfortran.dg/gomp/allocate-pinned-1.f90: Likewise. * gfortran.dg/gomp/allocate-4.f90: Likewise; add dg-error for previously missing diagnostic. * c-c++-common/gomp/allocate-18.c: New test. * c-c++-common/gomp/allocate-19.c: New test. * gfortran.dg/gomp/allocate-clause.f90: New test. * gfortran.dg/gomp/allocate-static-2.f90: New test. * gfortran.dg/gomp/allocate-static.f90: New test.
2024-10-07 16:45:14 +08:00
_Static_assert (GOMP_OMP_PREDEF_ALLOC_MAX == omp_thread_mem_alloc,
"GOMP_OMP_PREDEF_ALLOC_MAX == omp_thread_mem_alloc");
_Static_assert (GOMP_OMPX_PREDEF_ALLOC_MIN == ompx_gnu_min_predefined_alloc,
"GOMP_OMP_PREDEF_ALLOC_MAX == omp_thread_mem_alloc");
_Static_assert (GOMP_OMPX_PREDEF_ALLOC_MAX == ompx_gnu_max_predefined_alloc,
"GOMP_OMP_PREDEF_ALLOC_MAX == omp_thread_mem_alloc");
_Static_assert (GOMP_OMP_PREDEF_ALLOC_THREADS == omp_thread_mem_alloc,
"GOMP_OMP_PREDEF_ALLOC_THREADS == omp_thread_mem_alloc");
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
/* These macros may be overridden in config/<target>/allocator.c.
The defaults (no override) are to return NULL for pinned memory requests
and pass through to the regular OS calls otherwise.
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
The following definitions (ab)use comma operators to avoid unused
variable errors. */
#ifndef MEMSPACE_ALLOC
#define MEMSPACE_ALLOC(MEMSPACE, SIZE, PIN) \
(PIN ? NULL : malloc (((void)(MEMSPACE), (SIZE))))
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
#endif
#ifndef MEMSPACE_CALLOC
#define MEMSPACE_CALLOC(MEMSPACE, SIZE, PIN) \
(PIN ? NULL : calloc (1, (((void)(MEMSPACE), (SIZE)))))
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
#endif
#ifndef MEMSPACE_REALLOC
#define MEMSPACE_REALLOC(MEMSPACE, ADDR, OLDSIZE, SIZE, OLDPIN, PIN) \
((PIN) || (OLDPIN) ? NULL \
: realloc (ADDR, (((void)(MEMSPACE), (void)(OLDSIZE), (SIZE)))))
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
#endif
#ifndef MEMSPACE_FREE
#define MEMSPACE_FREE(MEMSPACE, ADDR, SIZE, PIN) \
if (PIN) free (((void)(MEMSPACE), (void)(SIZE), (ADDR)))
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
#endif
#ifndef MEMSPACE_VALIDATE
#define MEMSPACE_VALIDATE(MEMSPACE, ACCESS, PIN) \
(PIN ? 0 : ((void)(MEMSPACE), (void)(ACCESS), 1))
#endif
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
/* Map the predefined allocators to the correct memory space.
The index to this table is the omp_allocator_handle_t enum value.
When the user calls omp_alloc with a predefined allocator this
table determines what memory they get. */
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
static const omp_memspace_handle_t predefined_omp_alloc_mapping[] = {
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
omp_default_mem_space, /* omp_null_allocator doesn't actually use this. */
omp_default_mem_space, /* omp_default_mem_alloc. */
omp_large_cap_mem_space, /* omp_large_cap_mem_alloc. */
omp_const_mem_space, /* omp_const_mem_alloc. */
omp_high_bw_mem_space, /* omp_high_bw_mem_alloc. */
omp_low_lat_mem_space, /* omp_low_lat_mem_alloc. */
omp_low_lat_mem_space, /* omp_cgroup_mem_alloc (implementation defined). */
omp_low_lat_mem_space, /* omp_pteam_mem_alloc (implementation defined). */
omp_low_lat_mem_space, /* omp_thread_mem_alloc (implementation defined). */
};
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
static const omp_memspace_handle_t predefined_ompx_gnu_alloc_mapping[] = {
omp_default_mem_space, /* ompx_gnu_pinned_mem_alloc. */
};
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
#define ARRAY_SIZE(A) (sizeof (A) / sizeof ((A)[0]))
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
_Static_assert (ARRAY_SIZE (predefined_omp_alloc_mapping)
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
== omp_max_predefined_alloc + 1,
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
"predefined_omp_alloc_mapping must match omp_memspace_handle_t");
_Static_assert (ARRAY_SIZE (predefined_ompx_gnu_alloc_mapping)
== (ompx_gnu_max_predefined_alloc
- ompx_gnu_min_predefined_alloc) + 1,
"predefined_ompx_gnu_alloc_mapping must match"
" omp_memspace_handle_t");
static inline bool
predefined_allocator_p (omp_allocator_handle_t allocator)
{
return allocator <= ompx_gnu_max_predefined_alloc;
}
static inline omp_memspace_handle_t
predefined_alloc_mapping (omp_allocator_handle_t allocator)
{
if (allocator <= omp_max_predefined_alloc)
return predefined_omp_alloc_mapping[allocator];
else if (allocator >= ompx_gnu_min_predefined_alloc
&& allocator <= ompx_gnu_max_predefined_alloc)
{
int index = allocator - ompx_gnu_min_predefined_alloc;
return predefined_ompx_gnu_alloc_mapping[index];
}
else
/* This should never happen. */
return omp_default_mem_space;
}
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
enum gomp_numa_memkind_kind
{
GOMP_MEMKIND_NONE = 0,
#define GOMP_MEMKIND_KINDS \
GOMP_MEMKIND_KIND (HBW_INTERLEAVE), \
GOMP_MEMKIND_KIND (HBW_PREFERRED), \
GOMP_MEMKIND_KIND (DAX_KMEM_ALL), \
GOMP_MEMKIND_KIND (DAX_KMEM), \
GOMP_MEMKIND_KIND (INTERLEAVE), \
GOMP_MEMKIND_KIND (DEFAULT)
#define GOMP_MEMKIND_KIND(kind) GOMP_MEMKIND_##kind
GOMP_MEMKIND_KINDS,
#undef GOMP_MEMKIND_KIND
GOMP_MEMKIND_COUNT,
GOMP_MEMKIND_LIBNUMA = GOMP_MEMKIND_COUNT
};
struct omp_allocator_data
{
omp_memspace_handle_t memspace;
omp_uintptr_t alignment;
omp_uintptr_t pool_size;
omp_uintptr_t used_pool_size;
omp_allocator_handle_t fb_data;
unsigned int sync_hint : 8;
unsigned int access : 8;
unsigned int fallback : 8;
unsigned int pinned : 1;
unsigned int partition : 7;
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
unsigned int memkind : 8;
#endif
#ifndef HAVE_SYNC_BUILTINS
gomp_mutex_t lock;
#endif
};
struct omp_mem_header
{
void *ptr;
size_t size;
omp_allocator_handle_t allocator;
void *pad;
};
struct gomp_libnuma_data
{
void *numa_handle;
void *(*numa_alloc_local) (size_t);
void *(*numa_realloc) (void *, size_t, size_t);
void (*numa_free) (void *, size_t);
};
struct gomp_memkind_data
{
void *memkind_handle;
void *(*memkind_malloc) (void *, size_t);
void *(*memkind_calloc) (void *, size_t, size_t);
void *(*memkind_realloc) (void *, void *, size_t);
void (*memkind_free) (void *, void *);
int (*memkind_check_available) (void *);
void **kinds[GOMP_MEMKIND_COUNT];
};
#ifdef LIBGOMP_USE_LIBNUMA
static struct gomp_libnuma_data *libnuma_data;
static pthread_once_t libnuma_data_once = PTHREAD_ONCE_INIT;
static void
gomp_init_libnuma (void)
{
void *handle = dlopen ("libnuma.so.1", RTLD_LAZY);
struct gomp_libnuma_data *data;
data = calloc (1, sizeof (struct gomp_libnuma_data));
if (data == NULL)
{
if (handle)
dlclose (handle);
return;
}
if (handle)
{
int (*numa_available) (void);
numa_available
= (__typeof (numa_available)) dlsym (handle, "numa_available");
if (!numa_available || numa_available () != 0)
{
dlclose (handle);
handle = NULL;
}
}
if (!handle)
{
__atomic_store_n (&libnuma_data, data, MEMMODEL_RELEASE);
return;
}
data->numa_handle = handle;
data->numa_alloc_local
= (__typeof (data->numa_alloc_local)) dlsym (handle, "numa_alloc_local");
data->numa_realloc
= (__typeof (data->numa_realloc)) dlsym (handle, "numa_realloc");
data->numa_free
= (__typeof (data->numa_free)) dlsym (handle, "numa_free");
__atomic_store_n (&libnuma_data, data, MEMMODEL_RELEASE);
}
static struct gomp_libnuma_data *
gomp_get_libnuma (void)
{
struct gomp_libnuma_data *data
= __atomic_load_n (&libnuma_data, MEMMODEL_ACQUIRE);
if (data)
return data;
pthread_once (&libnuma_data_once, gomp_init_libnuma);
return __atomic_load_n (&libnuma_data, MEMMODEL_ACQUIRE);
}
#endif
#ifdef LIBGOMP_USE_MEMKIND
static struct gomp_memkind_data *memkind_data;
static pthread_once_t memkind_data_once = PTHREAD_ONCE_INIT;
static void
gomp_init_memkind (void)
{
void *handle = dlopen ("libmemkind.so.0", RTLD_LAZY);
struct gomp_memkind_data *data;
int i;
static const char *kinds[] = {
NULL,
#define GOMP_MEMKIND_KIND(kind) "MEMKIND_" #kind
GOMP_MEMKIND_KINDS
#undef GOMP_MEMKIND_KIND
};
data = calloc (1, sizeof (struct gomp_memkind_data));
if (data == NULL)
{
if (handle)
dlclose (handle);
return;
}
if (!handle)
{
__atomic_store_n (&memkind_data, data, MEMMODEL_RELEASE);
return;
}
data->memkind_handle = handle;
data->memkind_malloc
= (__typeof (data->memkind_malloc)) dlsym (handle, "memkind_malloc");
data->memkind_calloc
= (__typeof (data->memkind_calloc)) dlsym (handle, "memkind_calloc");
data->memkind_realloc
= (__typeof (data->memkind_realloc)) dlsym (handle, "memkind_realloc");
data->memkind_free
= (__typeof (data->memkind_free)) dlsym (handle, "memkind_free");
data->memkind_check_available
= (__typeof (data->memkind_check_available))
dlsym (handle, "memkind_check_available");
if (data->memkind_malloc
&& data->memkind_calloc
&& data->memkind_realloc
&& data->memkind_free
&& data->memkind_check_available)
for (i = 1; i < GOMP_MEMKIND_COUNT; ++i)
{
data->kinds[i] = (void **) dlsym (handle, kinds[i]);
if (data->kinds[i] && data->memkind_check_available (*data->kinds[i]))
data->kinds[i] = NULL;
}
__atomic_store_n (&memkind_data, data, MEMMODEL_RELEASE);
}
static struct gomp_memkind_data *
gomp_get_memkind (void)
{
struct gomp_memkind_data *data
= __atomic_load_n (&memkind_data, MEMMODEL_ACQUIRE);
if (data)
return data;
pthread_once (&memkind_data_once, gomp_init_memkind);
return __atomic_load_n (&memkind_data, MEMMODEL_ACQUIRE);
}
#endif
omp_allocator_handle_t
omp_init_allocator (omp_memspace_handle_t memspace, int ntraits,
const omp_alloctrait_t traits[])
{
struct omp_allocator_data data
= { memspace, 1, ~(uintptr_t) 0, 0, 0, omp_atv_contended, omp_atv_all,
omp_atv_default_mem_fb, omp_atv_false, omp_atv_environment,
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
GOMP_MEMKIND_NONE
#endif
};
struct omp_allocator_data *ret;
int i;
if (memspace > omp_low_lat_mem_space)
return omp_null_allocator;
for (i = 0; i < ntraits; i++)
switch (traits[i].key)
{
case omp_atk_sync_hint:
switch (traits[i].value)
{
case omp_atv_default:
data.sync_hint = omp_atv_contended;
break;
case omp_atv_contended:
case omp_atv_uncontended:
case omp_atv_serialized:
case omp_atv_private:
data.sync_hint = traits[i].value;
break;
default:
return omp_null_allocator;
}
break;
case omp_atk_alignment:
if (traits[i].value == omp_atv_default)
{
data.alignment = 1;
break;
}
if ((traits[i].value & (traits[i].value - 1)) != 0
|| !traits[i].value)
return omp_null_allocator;
data.alignment = traits[i].value;
break;
case omp_atk_access:
switch (traits[i].value)
{
case omp_atv_default:
data.access = omp_atv_all;
break;
case omp_atv_all:
case omp_atv_cgroup:
case omp_atv_pteam:
case omp_atv_thread:
data.access = traits[i].value;
break;
default:
return omp_null_allocator;
}
break;
case omp_atk_pool_size:
if (traits[i].value == omp_atv_default)
data.pool_size = ~(uintptr_t) 0;
else
data.pool_size = traits[i].value;
break;
case omp_atk_fallback:
switch (traits[i].value)
{
case omp_atv_default:
data.fallback = omp_atv_default_mem_fb;
break;
case omp_atv_default_mem_fb:
case omp_atv_null_fb:
case omp_atv_abort_fb:
case omp_atv_allocator_fb:
data.fallback = traits[i].value;
break;
default:
return omp_null_allocator;
}
break;
case omp_atk_fb_data:
data.fb_data = traits[i].value;
break;
case omp_atk_pinned:
switch (traits[i].value)
{
case omp_atv_default:
case omp_atv_false:
data.pinned = omp_atv_false;
break;
case omp_atv_true:
data.pinned = omp_atv_true;
break;
default:
return omp_null_allocator;
}
break;
case omp_atk_partition:
switch (traits[i].value)
{
case omp_atv_default:
data.partition = omp_atv_environment;
break;
case omp_atv_environment:
case omp_atv_nearest:
case omp_atv_blocked:
case omp_atv_interleaved:
data.partition = traits[i].value;
break;
default:
return omp_null_allocator;
}
break;
default:
return omp_null_allocator;
}
if (data.alignment < sizeof (void *))
data.alignment = sizeof (void *);
switch (memspace)
{
#ifdef LIBGOMP_USE_MEMKIND
case omp_high_bw_mem_space:
struct gomp_memkind_data *memkind_data;
memkind_data = gomp_get_memkind ();
if (data.partition == omp_atv_interleaved
&& memkind_data->kinds[GOMP_MEMKIND_HBW_INTERLEAVE])
{
data.memkind = GOMP_MEMKIND_HBW_INTERLEAVE;
break;
}
else if (memkind_data->kinds[GOMP_MEMKIND_HBW_PREFERRED])
{
data.memkind = GOMP_MEMKIND_HBW_PREFERRED;
break;
}
break;
case omp_large_cap_mem_space:
memkind_data = gomp_get_memkind ();
if (memkind_data->kinds[GOMP_MEMKIND_DAX_KMEM_ALL])
data.memkind = GOMP_MEMKIND_DAX_KMEM_ALL;
else if (memkind_data->kinds[GOMP_MEMKIND_DAX_KMEM])
data.memkind = GOMP_MEMKIND_DAX_KMEM;
break;
#endif
default:
#ifdef LIBGOMP_USE_MEMKIND
if (data.partition == omp_atv_interleaved)
{
memkind_data = gomp_get_memkind ();
if (memkind_data->kinds[GOMP_MEMKIND_INTERLEAVE])
data.memkind = GOMP_MEMKIND_INTERLEAVE;
}
#endif
break;
}
#ifdef LIBGOMP_USE_LIBNUMA
if (data.memkind == GOMP_MEMKIND_NONE && data.partition == omp_atv_nearest)
{
libnuma_data = gomp_get_libnuma ();
if (libnuma_data->numa_alloc_local != NULL)
data.memkind = GOMP_MEMKIND_LIBNUMA;
}
#endif
/* Reject unsupported memory spaces. */
if (!MEMSPACE_VALIDATE (data.memspace, data.access, data.pinned))
return omp_null_allocator;
ret = gomp_malloc (sizeof (struct omp_allocator_data));
*ret = data;
#ifndef HAVE_SYNC_BUILTINS
gomp_mutex_init (&ret->lock);
#endif
return (omp_allocator_handle_t) ret;
}
void
omp_destroy_allocator (omp_allocator_handle_t allocator)
{
if (allocator != omp_null_allocator)
{
#ifndef HAVE_SYNC_BUILTINS
gomp_mutex_destroy (&((struct omp_allocator_data *) allocator)->lock);
#endif
free ((void *) allocator);
}
}
ialias (omp_init_allocator)
ialias (omp_destroy_allocator)
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
void *
openmp: Implement allocate clause in omp lowering. For now, task/taskloop constructs aren't handled and C/C++ array reductions and reductions with task or inscan modifiers need further work. Instead of calling omp_alloc/omp_free (where the former doesn't have alignment argument and omp_aligned_alloc is 5.1 only feature), this calls GOMP_alloc/GOMP_free, so that the library can fail if it would fall back into NULL (exception is zero length allocations). 2020-11-12 Jakub Jelinek <jakub@redhat.com> gcc/ * builtin-types.def (BT_FN_PTR_SIZE_SIZE_PTRMODE): New function type. * omp-builtins.def (BUILT_IN_GOACC_DECLARE): Move earlier. (BUILT_IN_GOMP_ALLOC, BUILT_IN_GOMP_FREE): New builtins. * gimplify.c (gimplify_scan_omp_clauses): Force allocator into a decl if it is not NULL, INTEGER_CST or decl. (gimplify_adjust_omp_clauses): Clear GOVD_EXPLICIT on explicit clauses which are being removed. Remove allocate clauses for variables not seen if they are private, firstprivate or linear too. Call omp_notice_variable on the allocator otherwise. (gimplify_omp_for): Handle iterator vars mentioned in allocate clauses similarly to non-is_gimple_reg iterators. * omp-low.c (struct omp_context): Add allocate_map field. (delete_omp_context): Delete it. (scan_sharing_clauses): Fill it from allocate clauses. Remove it if mentioned also in shared clause. (lower_private_allocate): New function. (lower_rec_input_clauses): Handle allocate clause for privatized variables, except for task/taskloop, C/C++ array reductions for now and task/inscan variables. (lower_send_shared_vars): Don't consider variables in allocate_map as shared. * omp-expand.c (expand_omp_for_generic, expand_omp_for_static_nochunk, expand_omp_for_static_chunk): Use expand_omp_build_assign instead of gimple_build_assign + gsi_insert_after. * builtins.c (builtin_fnspec): Handle BUILTIN_GOMP_ALLOC and BUILTIN_GOMP_FREE. * tree-ssa-ccp.c (evaluate_stmt): Handle BUILTIN_GOMP_ALLOC. * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Handle BUILTIN_GOMP_ALLOC. (mark_all_reaching_defs_necessary_1): Handle BUILTIN_GOMP_ALLOC and BUILTIN_GOMP_FREE. (propagate_necessity): Likewise. gcc/fortran/ * f95-lang.c (ATTR_ALLOC_WARN_UNUSED_RESULT_SIZE_2_NOTHROW_LIST): Define. (gfc_init_builtin_functions): Add alloc_size and warn_unused_result attributes to __builtin_GOMP_alloc. * types.def (BT_PTRMODE): New primitive type. (BT_FN_VOID_PTR_PTRMODE, BT_FN_PTR_SIZE_SIZE_PTRMODE): New function types. libgomp/ * libgomp.map (GOMP_alloc, GOMP_free): Export at GOMP_5.0.1. * omp.h.in (omp_alloc): Add malloc and alloc_size attributes. * libgomp_g.h (GOMP_alloc, GOMP_free): Declare. * allocator.c (omp_aligned_alloc): New for now static function, add alignment argument and handle it. (omp_alloc): Reimplement using omp_aligned_alloc. (GOMP_alloc, GOMP_free): New functions. (omp_free): Add ialias. * testsuite/libgomp.c-c++-common/allocate-1.c: New test. * testsuite/libgomp.c++/allocate-1.C: New test.
2020-11-13 04:38:04 +08:00
omp_aligned_alloc (size_t alignment, size_t size,
omp_allocator_handle_t allocator)
{
struct omp_allocator_data *allocator_data;
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
size_t new_size, new_alignment;
void *ptr, *ret;
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
enum gomp_numa_memkind_kind memkind;
#endif
if (__builtin_expect (size == 0, 0))
return NULL;
retry:
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
new_alignment = alignment;
if (allocator == omp_null_allocator)
{
struct gomp_thread *thr = gomp_thread ();
if (thr->ts.def_allocator == omp_null_allocator)
thr->ts.def_allocator = gomp_def_allocator;
allocator = (omp_allocator_handle_t) thr->ts.def_allocator;
}
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
if (!predefined_allocator_p (allocator))
{
allocator_data = (struct omp_allocator_data *) allocator;
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
if (new_alignment < allocator_data->alignment)
new_alignment = allocator_data->alignment;
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
memkind = allocator_data->memkind;
#endif
}
else
{
allocator_data = NULL;
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
if (new_alignment < sizeof (void *))
new_alignment = sizeof (void *);
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
memkind = GOMP_MEMKIND_NONE;
#endif
#ifdef LIBGOMP_USE_MEMKIND
if (allocator == omp_high_bw_mem_alloc)
memkind = GOMP_MEMKIND_HBW_PREFERRED;
else if (allocator == omp_large_cap_mem_alloc)
memkind = GOMP_MEMKIND_DAX_KMEM_ALL;
if (memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
if (!memkind_data->kinds[memkind])
memkind = GOMP_MEMKIND_NONE;
}
#endif
}
new_size = sizeof (struct omp_mem_header);
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
if (new_alignment > sizeof (void *))
new_size += new_alignment - sizeof (void *);
if (__builtin_add_overflow (size, new_size, &new_size))
goto fail;
#ifdef OMP_LOW_LAT_MEM_ALLOC_INVALID
if (allocator == omp_low_lat_mem_alloc)
goto fail;
#endif
if (__builtin_expect (allocator_data
&& allocator_data->pool_size < ~(uintptr_t) 0, 0))
{
uintptr_t used_pool_size;
if (new_size > allocator_data->pool_size)
goto fail;
#ifdef HAVE_SYNC_BUILTINS
used_pool_size = __atomic_load_n (&allocator_data->used_pool_size,
MEMMODEL_RELAXED);
do
{
uintptr_t new_pool_size;
if (__builtin_add_overflow (used_pool_size, new_size,
&new_pool_size)
|| new_pool_size > allocator_data->pool_size)
goto fail;
if (__atomic_compare_exchange_n (&allocator_data->used_pool_size,
&used_pool_size, new_pool_size,
true, MEMMODEL_RELAXED,
MEMMODEL_RELAXED))
break;
}
while (1);
#else
gomp_mutex_lock (&allocator_data->lock);
if (__builtin_add_overflow (allocator_data->used_pool_size, new_size,
&used_pool_size)
|| used_pool_size > allocator_data->pool_size)
{
gomp_mutex_unlock (&allocator_data->lock);
goto fail;
}
allocator_data->used_pool_size = used_pool_size;
gomp_mutex_unlock (&allocator_data->lock);
#endif
#ifdef LIBGOMP_USE_LIBNUMA
if (memkind == GOMP_MEMKIND_LIBNUMA)
ptr = libnuma_data->numa_alloc_local (new_size);
# ifdef LIBGOMP_USE_MEMKIND
else
# endif
#endif
#ifdef LIBGOMP_USE_MEMKIND
if (memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
void *kind = *memkind_data->kinds[memkind];
ptr = memkind_data->memkind_malloc (kind, new_size);
}
else
#endif
ptr = MEMSPACE_ALLOC (allocator_data->memspace, new_size,
allocator_data->pinned);
if (ptr == NULL)
{
#ifdef HAVE_SYNC_BUILTINS
__atomic_add_fetch (&allocator_data->used_pool_size, -new_size,
MEMMODEL_RELAXED);
#else
gomp_mutex_lock (&allocator_data->lock);
allocator_data->used_pool_size -= new_size;
gomp_mutex_unlock (&allocator_data->lock);
#endif
goto fail;
}
}
else
{
#ifdef LIBGOMP_USE_LIBNUMA
if (memkind == GOMP_MEMKIND_LIBNUMA)
ptr = libnuma_data->numa_alloc_local (new_size);
# ifdef LIBGOMP_USE_MEMKIND
else
# endif
#endif
#ifdef LIBGOMP_USE_MEMKIND
if (memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
void *kind = *memkind_data->kinds[memkind];
ptr = memkind_data->memkind_malloc (kind, new_size);
}
else
#endif
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
{
omp_memspace_handle_t memspace;
memspace = (allocator_data
? allocator_data->memspace
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
: predefined_alloc_mapping (allocator));
int pinned = (allocator_data
? allocator_data->pinned
: allocator == ompx_gnu_pinned_mem_alloc);
ptr = MEMSPACE_ALLOC (memspace, new_size, pinned);
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
}
if (ptr == NULL)
goto fail;
}
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
if (new_alignment > sizeof (void *))
ret = (void *) (((uintptr_t) ptr
+ sizeof (struct omp_mem_header)
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
+ new_alignment - sizeof (void *))
& ~(new_alignment - 1));
else
ret = (char *) ptr + sizeof (struct omp_mem_header);
((struct omp_mem_header *) ret)[-1].ptr = ptr;
((struct omp_mem_header *) ret)[-1].size = new_size;
((struct omp_mem_header *) ret)[-1].allocator = allocator;
return ret;
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
fail:;
int fallback = (allocator_data
? allocator_data->fallback
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
: (allocator == omp_default_mem_alloc
|| allocator == ompx_gnu_pinned_mem_alloc)
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
? omp_atv_null_fb
: omp_atv_default_mem_fb);
switch (fallback)
{
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
case omp_atv_default_mem_fb:
allocator = omp_default_mem_alloc;
goto retry;
case omp_atv_null_fb:
break;
default:
case omp_atv_abort_fb:
gomp_fatal ("Out of memory allocating %lu bytes",
(unsigned long) size);
case omp_atv_allocator_fb:
allocator = allocator_data->fb_data;
goto retry;
}
return NULL;
}
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
ialias (omp_aligned_alloc)
openmp: Implement allocate clause in omp lowering. For now, task/taskloop constructs aren't handled and C/C++ array reductions and reductions with task or inscan modifiers need further work. Instead of calling omp_alloc/omp_free (where the former doesn't have alignment argument and omp_aligned_alloc is 5.1 only feature), this calls GOMP_alloc/GOMP_free, so that the library can fail if it would fall back into NULL (exception is zero length allocations). 2020-11-12 Jakub Jelinek <jakub@redhat.com> gcc/ * builtin-types.def (BT_FN_PTR_SIZE_SIZE_PTRMODE): New function type. * omp-builtins.def (BUILT_IN_GOACC_DECLARE): Move earlier. (BUILT_IN_GOMP_ALLOC, BUILT_IN_GOMP_FREE): New builtins. * gimplify.c (gimplify_scan_omp_clauses): Force allocator into a decl if it is not NULL, INTEGER_CST or decl. (gimplify_adjust_omp_clauses): Clear GOVD_EXPLICIT on explicit clauses which are being removed. Remove allocate clauses for variables not seen if they are private, firstprivate or linear too. Call omp_notice_variable on the allocator otherwise. (gimplify_omp_for): Handle iterator vars mentioned in allocate clauses similarly to non-is_gimple_reg iterators. * omp-low.c (struct omp_context): Add allocate_map field. (delete_omp_context): Delete it. (scan_sharing_clauses): Fill it from allocate clauses. Remove it if mentioned also in shared clause. (lower_private_allocate): New function. (lower_rec_input_clauses): Handle allocate clause for privatized variables, except for task/taskloop, C/C++ array reductions for now and task/inscan variables. (lower_send_shared_vars): Don't consider variables in allocate_map as shared. * omp-expand.c (expand_omp_for_generic, expand_omp_for_static_nochunk, expand_omp_for_static_chunk): Use expand_omp_build_assign instead of gimple_build_assign + gsi_insert_after. * builtins.c (builtin_fnspec): Handle BUILTIN_GOMP_ALLOC and BUILTIN_GOMP_FREE. * tree-ssa-ccp.c (evaluate_stmt): Handle BUILTIN_GOMP_ALLOC. * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Handle BUILTIN_GOMP_ALLOC. (mark_all_reaching_defs_necessary_1): Handle BUILTIN_GOMP_ALLOC and BUILTIN_GOMP_FREE. (propagate_necessity): Likewise. gcc/fortran/ * f95-lang.c (ATTR_ALLOC_WARN_UNUSED_RESULT_SIZE_2_NOTHROW_LIST): Define. (gfc_init_builtin_functions): Add alloc_size and warn_unused_result attributes to __builtin_GOMP_alloc. * types.def (BT_PTRMODE): New primitive type. (BT_FN_VOID_PTR_PTRMODE, BT_FN_PTR_SIZE_SIZE_PTRMODE): New function types. libgomp/ * libgomp.map (GOMP_alloc, GOMP_free): Export at GOMP_5.0.1. * omp.h.in (omp_alloc): Add malloc and alloc_size attributes. * libgomp_g.h (GOMP_alloc, GOMP_free): Declare. * allocator.c (omp_aligned_alloc): New for now static function, add alignment argument and handle it. (omp_alloc): Reimplement using omp_aligned_alloc. (GOMP_alloc, GOMP_free): New functions. (omp_free): Add ialias. * testsuite/libgomp.c-c++-common/allocate-1.c: New test. * testsuite/libgomp.c++/allocate-1.C: New test.
2020-11-13 04:38:04 +08:00
void *
omp_alloc (size_t size, omp_allocator_handle_t allocator)
{
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
return ialias_call (omp_aligned_alloc) (1, size, allocator);
openmp: Implement allocate clause in omp lowering. For now, task/taskloop constructs aren't handled and C/C++ array reductions and reductions with task or inscan modifiers need further work. Instead of calling omp_alloc/omp_free (where the former doesn't have alignment argument and omp_aligned_alloc is 5.1 only feature), this calls GOMP_alloc/GOMP_free, so that the library can fail if it would fall back into NULL (exception is zero length allocations). 2020-11-12 Jakub Jelinek <jakub@redhat.com> gcc/ * builtin-types.def (BT_FN_PTR_SIZE_SIZE_PTRMODE): New function type. * omp-builtins.def (BUILT_IN_GOACC_DECLARE): Move earlier. (BUILT_IN_GOMP_ALLOC, BUILT_IN_GOMP_FREE): New builtins. * gimplify.c (gimplify_scan_omp_clauses): Force allocator into a decl if it is not NULL, INTEGER_CST or decl. (gimplify_adjust_omp_clauses): Clear GOVD_EXPLICIT on explicit clauses which are being removed. Remove allocate clauses for variables not seen if they are private, firstprivate or linear too. Call omp_notice_variable on the allocator otherwise. (gimplify_omp_for): Handle iterator vars mentioned in allocate clauses similarly to non-is_gimple_reg iterators. * omp-low.c (struct omp_context): Add allocate_map field. (delete_omp_context): Delete it. (scan_sharing_clauses): Fill it from allocate clauses. Remove it if mentioned also in shared clause. (lower_private_allocate): New function. (lower_rec_input_clauses): Handle allocate clause for privatized variables, except for task/taskloop, C/C++ array reductions for now and task/inscan variables. (lower_send_shared_vars): Don't consider variables in allocate_map as shared. * omp-expand.c (expand_omp_for_generic, expand_omp_for_static_nochunk, expand_omp_for_static_chunk): Use expand_omp_build_assign instead of gimple_build_assign + gsi_insert_after. * builtins.c (builtin_fnspec): Handle BUILTIN_GOMP_ALLOC and BUILTIN_GOMP_FREE. * tree-ssa-ccp.c (evaluate_stmt): Handle BUILTIN_GOMP_ALLOC. * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Handle BUILTIN_GOMP_ALLOC. (mark_all_reaching_defs_necessary_1): Handle BUILTIN_GOMP_ALLOC and BUILTIN_GOMP_FREE. (propagate_necessity): Likewise. gcc/fortran/ * f95-lang.c (ATTR_ALLOC_WARN_UNUSED_RESULT_SIZE_2_NOTHROW_LIST): Define. (gfc_init_builtin_functions): Add alloc_size and warn_unused_result attributes to __builtin_GOMP_alloc. * types.def (BT_PTRMODE): New primitive type. (BT_FN_VOID_PTR_PTRMODE, BT_FN_PTR_SIZE_SIZE_PTRMODE): New function types. libgomp/ * libgomp.map (GOMP_alloc, GOMP_free): Export at GOMP_5.0.1. * omp.h.in (omp_alloc): Add malloc and alloc_size attributes. * libgomp_g.h (GOMP_alloc, GOMP_free): Declare. * allocator.c (omp_aligned_alloc): New for now static function, add alignment argument and handle it. (omp_alloc): Reimplement using omp_aligned_alloc. (GOMP_alloc, GOMP_free): New functions. (omp_free): Add ialias. * testsuite/libgomp.c-c++-common/allocate-1.c: New test. * testsuite/libgomp.c++/allocate-1.C: New test.
2020-11-13 04:38:04 +08:00
}
/* Like omp_aligned_alloc, but apply on top of that:
"For allocations that arise from this ... the null_fb value of the
fallback allocator trait behaves as if the abort_fb had been specified." */
void *
GOMP_alloc (size_t alignment, size_t size, uintptr_t allocator)
{
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
void *ret
= ialias_call (omp_aligned_alloc) (alignment, size,
(omp_allocator_handle_t) allocator);
openmp: Implement allocate clause in omp lowering. For now, task/taskloop constructs aren't handled and C/C++ array reductions and reductions with task or inscan modifiers need further work. Instead of calling omp_alloc/omp_free (where the former doesn't have alignment argument and omp_aligned_alloc is 5.1 only feature), this calls GOMP_alloc/GOMP_free, so that the library can fail if it would fall back into NULL (exception is zero length allocations). 2020-11-12 Jakub Jelinek <jakub@redhat.com> gcc/ * builtin-types.def (BT_FN_PTR_SIZE_SIZE_PTRMODE): New function type. * omp-builtins.def (BUILT_IN_GOACC_DECLARE): Move earlier. (BUILT_IN_GOMP_ALLOC, BUILT_IN_GOMP_FREE): New builtins. * gimplify.c (gimplify_scan_omp_clauses): Force allocator into a decl if it is not NULL, INTEGER_CST or decl. (gimplify_adjust_omp_clauses): Clear GOVD_EXPLICIT on explicit clauses which are being removed. Remove allocate clauses for variables not seen if they are private, firstprivate or linear too. Call omp_notice_variable on the allocator otherwise. (gimplify_omp_for): Handle iterator vars mentioned in allocate clauses similarly to non-is_gimple_reg iterators. * omp-low.c (struct omp_context): Add allocate_map field. (delete_omp_context): Delete it. (scan_sharing_clauses): Fill it from allocate clauses. Remove it if mentioned also in shared clause. (lower_private_allocate): New function. (lower_rec_input_clauses): Handle allocate clause for privatized variables, except for task/taskloop, C/C++ array reductions for now and task/inscan variables. (lower_send_shared_vars): Don't consider variables in allocate_map as shared. * omp-expand.c (expand_omp_for_generic, expand_omp_for_static_nochunk, expand_omp_for_static_chunk): Use expand_omp_build_assign instead of gimple_build_assign + gsi_insert_after. * builtins.c (builtin_fnspec): Handle BUILTIN_GOMP_ALLOC and BUILTIN_GOMP_FREE. * tree-ssa-ccp.c (evaluate_stmt): Handle BUILTIN_GOMP_ALLOC. * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Handle BUILTIN_GOMP_ALLOC. (mark_all_reaching_defs_necessary_1): Handle BUILTIN_GOMP_ALLOC and BUILTIN_GOMP_FREE. (propagate_necessity): Likewise. gcc/fortran/ * f95-lang.c (ATTR_ALLOC_WARN_UNUSED_RESULT_SIZE_2_NOTHROW_LIST): Define. (gfc_init_builtin_functions): Add alloc_size and warn_unused_result attributes to __builtin_GOMP_alloc. * types.def (BT_PTRMODE): New primitive type. (BT_FN_VOID_PTR_PTRMODE, BT_FN_PTR_SIZE_SIZE_PTRMODE): New function types. libgomp/ * libgomp.map (GOMP_alloc, GOMP_free): Export at GOMP_5.0.1. * omp.h.in (omp_alloc): Add malloc and alloc_size attributes. * libgomp_g.h (GOMP_alloc, GOMP_free): Declare. * allocator.c (omp_aligned_alloc): New for now static function, add alignment argument and handle it. (omp_alloc): Reimplement using omp_aligned_alloc. (GOMP_alloc, GOMP_free): New functions. (omp_free): Add ialias. * testsuite/libgomp.c-c++-common/allocate-1.c: New test. * testsuite/libgomp.c++/allocate-1.C: New test.
2020-11-13 04:38:04 +08:00
if (__builtin_expect (ret == NULL, 0) && size)
gomp_fatal ("Out of memory allocating %lu bytes",
(unsigned long) size);
return ret;
}
void
omp_free (void *ptr, omp_allocator_handle_t allocator)
{
struct omp_mem_header *data;
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
omp_memspace_handle_t memspace = omp_default_mem_space;
int pinned = false;
if (ptr == NULL)
return;
(void) allocator;
data = &((struct omp_mem_header *) ptr)[-1];
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
if (!predefined_allocator_p (data->allocator))
{
struct omp_allocator_data *allocator_data
= (struct omp_allocator_data *) (data->allocator);
if (allocator_data->pool_size < ~(uintptr_t) 0)
{
#ifdef HAVE_SYNC_BUILTINS
__atomic_add_fetch (&allocator_data->used_pool_size, -data->size,
MEMMODEL_RELAXED);
#else
gomp_mutex_lock (&allocator_data->lock);
allocator_data->used_pool_size -= data->size;
gomp_mutex_unlock (&allocator_data->lock);
#endif
}
#ifdef LIBGOMP_USE_LIBNUMA
if (allocator_data->memkind == GOMP_MEMKIND_LIBNUMA)
{
libnuma_data->numa_free (data->ptr, data->size);
return;
}
# ifdef LIBGOMP_USE_MEMKIND
else
# endif
#endif
#ifdef LIBGOMP_USE_MEMKIND
if (allocator_data->memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
void *kind = *memkind_data->kinds[allocator_data->memkind];
memkind_data->memkind_free (kind, data->ptr);
return;
}
#endif
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
memspace = allocator_data->memspace;
pinned = allocator_data->pinned;
}
else
{
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
#ifdef LIBGOMP_USE_MEMKIND
enum gomp_numa_memkind_kind memkind = GOMP_MEMKIND_NONE;
if (data->allocator == omp_high_bw_mem_alloc)
memkind = GOMP_MEMKIND_HBW_PREFERRED;
else if (data->allocator == omp_large_cap_mem_alloc)
memkind = GOMP_MEMKIND_DAX_KMEM_ALL;
if (memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
if (memkind_data->kinds[memkind])
{
void *kind = *memkind_data->kinds[memkind];
memkind_data->memkind_free (kind, data->ptr);
return;
}
}
#endif
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
memspace = predefined_alloc_mapping (data->allocator);
pinned = (data->allocator == ompx_gnu_pinned_mem_alloc);
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
}
MEMSPACE_FREE (memspace, data->ptr, data->size, pinned);
}
openmp: Implement allocate clause in omp lowering. For now, task/taskloop constructs aren't handled and C/C++ array reductions and reductions with task or inscan modifiers need further work. Instead of calling omp_alloc/omp_free (where the former doesn't have alignment argument and omp_aligned_alloc is 5.1 only feature), this calls GOMP_alloc/GOMP_free, so that the library can fail if it would fall back into NULL (exception is zero length allocations). 2020-11-12 Jakub Jelinek <jakub@redhat.com> gcc/ * builtin-types.def (BT_FN_PTR_SIZE_SIZE_PTRMODE): New function type. * omp-builtins.def (BUILT_IN_GOACC_DECLARE): Move earlier. (BUILT_IN_GOMP_ALLOC, BUILT_IN_GOMP_FREE): New builtins. * gimplify.c (gimplify_scan_omp_clauses): Force allocator into a decl if it is not NULL, INTEGER_CST or decl. (gimplify_adjust_omp_clauses): Clear GOVD_EXPLICIT on explicit clauses which are being removed. Remove allocate clauses for variables not seen if they are private, firstprivate or linear too. Call omp_notice_variable on the allocator otherwise. (gimplify_omp_for): Handle iterator vars mentioned in allocate clauses similarly to non-is_gimple_reg iterators. * omp-low.c (struct omp_context): Add allocate_map field. (delete_omp_context): Delete it. (scan_sharing_clauses): Fill it from allocate clauses. Remove it if mentioned also in shared clause. (lower_private_allocate): New function. (lower_rec_input_clauses): Handle allocate clause for privatized variables, except for task/taskloop, C/C++ array reductions for now and task/inscan variables. (lower_send_shared_vars): Don't consider variables in allocate_map as shared. * omp-expand.c (expand_omp_for_generic, expand_omp_for_static_nochunk, expand_omp_for_static_chunk): Use expand_omp_build_assign instead of gimple_build_assign + gsi_insert_after. * builtins.c (builtin_fnspec): Handle BUILTIN_GOMP_ALLOC and BUILTIN_GOMP_FREE. * tree-ssa-ccp.c (evaluate_stmt): Handle BUILTIN_GOMP_ALLOC. * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Handle BUILTIN_GOMP_ALLOC. (mark_all_reaching_defs_necessary_1): Handle BUILTIN_GOMP_ALLOC and BUILTIN_GOMP_FREE. (propagate_necessity): Likewise. gcc/fortran/ * f95-lang.c (ATTR_ALLOC_WARN_UNUSED_RESULT_SIZE_2_NOTHROW_LIST): Define. (gfc_init_builtin_functions): Add alloc_size and warn_unused_result attributes to __builtin_GOMP_alloc. * types.def (BT_PTRMODE): New primitive type. (BT_FN_VOID_PTR_PTRMODE, BT_FN_PTR_SIZE_SIZE_PTRMODE): New function types. libgomp/ * libgomp.map (GOMP_alloc, GOMP_free): Export at GOMP_5.0.1. * omp.h.in (omp_alloc): Add malloc and alloc_size attributes. * libgomp_g.h (GOMP_alloc, GOMP_free): Declare. * allocator.c (omp_aligned_alloc): New for now static function, add alignment argument and handle it. (omp_alloc): Reimplement using omp_aligned_alloc. (GOMP_alloc, GOMP_free): New functions. (omp_free): Add ialias. * testsuite/libgomp.c-c++-common/allocate-1.c: New test. * testsuite/libgomp.c++/allocate-1.C: New test.
2020-11-13 04:38:04 +08:00
ialias (omp_free)
void
GOMP_free (void *ptr, uintptr_t allocator)
{
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
return ialias_call (omp_free) (ptr, (omp_allocator_handle_t) allocator);
}
void *
omp_aligned_calloc (size_t alignment, size_t nmemb, size_t size,
omp_allocator_handle_t allocator)
{
struct omp_allocator_data *allocator_data;
size_t new_size, size_temp, new_alignment;
void *ptr, *ret;
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
enum gomp_numa_memkind_kind memkind;
#endif
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
if (__builtin_expect (size == 0 || nmemb == 0, 0))
return NULL;
retry:
new_alignment = alignment;
if (allocator == omp_null_allocator)
{
struct gomp_thread *thr = gomp_thread ();
if (thr->ts.def_allocator == omp_null_allocator)
thr->ts.def_allocator = gomp_def_allocator;
allocator = (omp_allocator_handle_t) thr->ts.def_allocator;
}
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
if (!predefined_allocator_p (allocator))
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
{
allocator_data = (struct omp_allocator_data *) allocator;
if (new_alignment < allocator_data->alignment)
new_alignment = allocator_data->alignment;
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
memkind = allocator_data->memkind;
#endif
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
}
else
{
allocator_data = NULL;
if (new_alignment < sizeof (void *))
new_alignment = sizeof (void *);
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
memkind = GOMP_MEMKIND_NONE;
#endif
#ifdef LIBGOMP_USE_MEMKIND
if (allocator == omp_high_bw_mem_alloc)
memkind = GOMP_MEMKIND_HBW_PREFERRED;
else if (allocator == omp_large_cap_mem_alloc)
memkind = GOMP_MEMKIND_DAX_KMEM_ALL;
if (memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
if (!memkind_data->kinds[memkind])
memkind = GOMP_MEMKIND_NONE;
}
#endif
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
}
new_size = sizeof (struct omp_mem_header);
if (new_alignment > sizeof (void *))
new_size += new_alignment - sizeof (void *);
if (__builtin_mul_overflow (size, nmemb, &size_temp))
goto fail;
if (__builtin_add_overflow (size_temp, new_size, &new_size))
goto fail;
#ifdef OMP_LOW_LAT_MEM_ALLOC_INVALID
if (allocator == omp_low_lat_mem_alloc)
goto fail;
#endif
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
if (__builtin_expect (allocator_data
&& allocator_data->pool_size < ~(uintptr_t) 0, 0))
{
uintptr_t used_pool_size;
if (new_size > allocator_data->pool_size)
goto fail;
#ifdef HAVE_SYNC_BUILTINS
used_pool_size = __atomic_load_n (&allocator_data->used_pool_size,
MEMMODEL_RELAXED);
do
{
uintptr_t new_pool_size;
if (__builtin_add_overflow (used_pool_size, new_size,
&new_pool_size)
|| new_pool_size > allocator_data->pool_size)
goto fail;
if (__atomic_compare_exchange_n (&allocator_data->used_pool_size,
&used_pool_size, new_pool_size,
true, MEMMODEL_RELAXED,
MEMMODEL_RELAXED))
break;
}
while (1);
#else
gomp_mutex_lock (&allocator_data->lock);
if (__builtin_add_overflow (allocator_data->used_pool_size, new_size,
&used_pool_size)
|| used_pool_size > allocator_data->pool_size)
{
gomp_mutex_unlock (&allocator_data->lock);
goto fail;
}
allocator_data->used_pool_size = used_pool_size;
gomp_mutex_unlock (&allocator_data->lock);
#endif
#ifdef LIBGOMP_USE_LIBNUMA
if (memkind == GOMP_MEMKIND_LIBNUMA)
/* numa_alloc_local uses mmap with MAP_ANONYMOUS, returning
memory that is initialized to zero. */
ptr = libnuma_data->numa_alloc_local (new_size);
# ifdef LIBGOMP_USE_MEMKIND
else
# endif
#endif
#ifdef LIBGOMP_USE_MEMKIND
if (memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
void *kind = *memkind_data->kinds[memkind];
ptr = memkind_data->memkind_calloc (kind, 1, new_size);
}
else
#endif
ptr = MEMSPACE_CALLOC (allocator_data->memspace, new_size,
allocator_data->pinned);
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
if (ptr == NULL)
{
#ifdef HAVE_SYNC_BUILTINS
__atomic_add_fetch (&allocator_data->used_pool_size, -new_size,
MEMMODEL_RELAXED);
#else
gomp_mutex_lock (&allocator_data->lock);
allocator_data->used_pool_size -= new_size;
gomp_mutex_unlock (&allocator_data->lock);
#endif
goto fail;
}
}
else
{
#ifdef LIBGOMP_USE_LIBNUMA
if (memkind == GOMP_MEMKIND_LIBNUMA)
/* numa_alloc_local uses mmap with MAP_ANONYMOUS, returning
memory that is initialized to zero. */
ptr = libnuma_data->numa_alloc_local (new_size);
# ifdef LIBGOMP_USE_MEMKIND
else
# endif
#endif
#ifdef LIBGOMP_USE_MEMKIND
if (memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
void *kind = *memkind_data->kinds[memkind];
ptr = memkind_data->memkind_calloc (kind, 1, new_size);
}
else
#endif
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
{
omp_memspace_handle_t memspace;
memspace = (allocator_data
? allocator_data->memspace
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
: predefined_alloc_mapping (allocator));
int pinned = (allocator_data
? allocator_data->pinned
: allocator == ompx_gnu_pinned_mem_alloc);
ptr = MEMSPACE_CALLOC (memspace, new_size, pinned);
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
}
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
if (ptr == NULL)
goto fail;
}
if (new_alignment > sizeof (void *))
ret = (void *) (((uintptr_t) ptr
+ sizeof (struct omp_mem_header)
+ new_alignment - sizeof (void *))
& ~(new_alignment - 1));
else
ret = (char *) ptr + sizeof (struct omp_mem_header);
((struct omp_mem_header *) ret)[-1].ptr = ptr;
((struct omp_mem_header *) ret)[-1].size = new_size;
((struct omp_mem_header *) ret)[-1].allocator = allocator;
return ret;
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
fail:;
int fallback = (allocator_data
? allocator_data->fallback
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
: (allocator == omp_default_mem_alloc
|| allocator == ompx_gnu_pinned_mem_alloc)
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
? omp_atv_null_fb
: omp_atv_default_mem_fb);
switch (fallback)
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
{
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
case omp_atv_default_mem_fb:
allocator = omp_default_mem_alloc;
goto retry;
case omp_atv_null_fb:
break;
default:
case omp_atv_abort_fb:
gomp_fatal ("Out of memory allocating %lu bytes",
(unsigned long) (size * nmemb));
case omp_atv_allocator_fb:
allocator = allocator_data->fb_data;
goto retry;
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
}
return NULL;
}
ialias (omp_aligned_calloc)
void *
omp_calloc (size_t nmemb, size_t size, omp_allocator_handle_t allocator)
{
return ialias_call (omp_aligned_calloc) (1, nmemb, size, allocator);
}
void *
omp_realloc (void *ptr, size_t size, omp_allocator_handle_t allocator,
omp_allocator_handle_t free_allocator)
{
struct omp_allocator_data *allocator_data, *free_allocator_data;
size_t new_size, old_size, new_alignment, old_alignment;
void *new_ptr, *ret;
struct omp_mem_header *data;
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
enum gomp_numa_memkind_kind memkind, free_memkind;
#endif
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
if (__builtin_expect (ptr == NULL, 0))
return ialias_call (omp_aligned_alloc) (1, size, allocator);
if (__builtin_expect (size == 0, 0))
{
ialias_call (omp_free) (ptr, free_allocator);
return NULL;
}
data = &((struct omp_mem_header *) ptr)[-1];
free_allocator = data->allocator;
retry:
new_alignment = sizeof (void *);
if (allocator == omp_null_allocator)
allocator = free_allocator;
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
if (!predefined_allocator_p (allocator))
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
{
allocator_data = (struct omp_allocator_data *) allocator;
if (new_alignment < allocator_data->alignment)
new_alignment = allocator_data->alignment;
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
memkind = allocator_data->memkind;
#endif
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
}
else
{
allocator_data = NULL;
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
memkind = GOMP_MEMKIND_NONE;
#endif
#ifdef LIBGOMP_USE_MEMKIND
if (allocator == omp_high_bw_mem_alloc)
memkind = GOMP_MEMKIND_HBW_PREFERRED;
else if (allocator == omp_large_cap_mem_alloc)
memkind = GOMP_MEMKIND_DAX_KMEM_ALL;
if (memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
if (!memkind_data->kinds[memkind])
memkind = GOMP_MEMKIND_NONE;
}
#endif
}
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
if (!predefined_allocator_p (free_allocator))
{
free_allocator_data = (struct omp_allocator_data *) free_allocator;
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
free_memkind = free_allocator_data->memkind;
#endif
}
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
else
{
free_allocator_data = NULL;
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
free_memkind = GOMP_MEMKIND_NONE;
#endif
#ifdef LIBGOMP_USE_MEMKIND
if (free_allocator == omp_high_bw_mem_alloc)
free_memkind = GOMP_MEMKIND_HBW_PREFERRED;
else if (free_allocator == omp_large_cap_mem_alloc)
free_memkind = GOMP_MEMKIND_DAX_KMEM_ALL;
if (free_memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
if (!memkind_data->kinds[free_memkind])
free_memkind = GOMP_MEMKIND_NONE;
}
#endif
}
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
old_alignment = (uintptr_t) ptr - (uintptr_t) (data->ptr);
new_size = sizeof (struct omp_mem_header);
if (new_alignment > sizeof (void *))
new_size += new_alignment - sizeof (void *);
if (__builtin_add_overflow (size, new_size, &new_size))
goto fail;
old_size = data->size;
#ifdef OMP_LOW_LAT_MEM_ALLOC_INVALID
if (allocator == omp_low_lat_mem_alloc)
goto fail;
#endif
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
if (__builtin_expect (allocator_data
&& allocator_data->pool_size < ~(uintptr_t) 0, 0))
{
uintptr_t used_pool_size;
size_t prev_size = 0;
/* Check if we can use realloc. Don't use it if extra alignment
was used previously or newly, because realloc might return a pointer
with different alignment and then we'd need to memmove the data
again. */
if (free_allocator_data
&& free_allocator_data == allocator_data
&& new_alignment == sizeof (void *)
&& old_alignment == sizeof (struct omp_mem_header))
prev_size = old_size;
if (new_size > prev_size
&& new_size - prev_size > allocator_data->pool_size)
goto fail;
#ifdef HAVE_SYNC_BUILTINS
used_pool_size = __atomic_load_n (&allocator_data->used_pool_size,
MEMMODEL_RELAXED);
do
{
uintptr_t new_pool_size;
if (new_size > prev_size)
{
if (__builtin_add_overflow (used_pool_size, new_size - prev_size,
&new_pool_size)
|| new_pool_size > allocator_data->pool_size)
goto fail;
}
else
new_pool_size = used_pool_size + new_size - prev_size;
if (__atomic_compare_exchange_n (&allocator_data->used_pool_size,
&used_pool_size, new_pool_size,
true, MEMMODEL_RELAXED,
MEMMODEL_RELAXED))
break;
}
while (1);
#else
gomp_mutex_lock (&allocator_data->lock);
if (new_size > prev_size)
{
if (__builtin_add_overflow (allocator_data->used_pool_size,
new_size - prev_size,
&used_pool_size)
|| used_pool_size > allocator_data->pool_size)
{
gomp_mutex_unlock (&allocator_data->lock);
goto fail;
}
}
else
used_pool_size = (allocator_data->used_pool_size
+ new_size - prev_size);
allocator_data->used_pool_size = used_pool_size;
gomp_mutex_unlock (&allocator_data->lock);
#endif
#ifdef LIBGOMP_USE_LIBNUMA
if (memkind == GOMP_MEMKIND_LIBNUMA)
{
if (prev_size)
new_ptr = libnuma_data->numa_realloc (data->ptr, data->size,
new_size);
else
new_ptr = libnuma_data->numa_alloc_local (new_size);
}
# ifdef LIBGOMP_USE_MEMKIND
else
# endif
#endif
#ifdef LIBGOMP_USE_MEMKIND
if (memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
void *kind = *memkind_data->kinds[memkind];
if (prev_size)
new_ptr = memkind_data->memkind_realloc (kind, data->ptr,
new_size);
else
new_ptr = memkind_data->memkind_malloc (kind, new_size);
}
else
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
#endif
if (prev_size)
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
{
int was_pinned = (free_allocator_data
? free_allocator_data->pinned
: free_allocator == ompx_gnu_pinned_mem_alloc);
new_ptr = MEMSPACE_REALLOC (allocator_data->memspace, data->ptr,
data->size, new_size, was_pinned,
allocator_data->pinned);
}
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
else
new_ptr = MEMSPACE_ALLOC (allocator_data->memspace, new_size,
allocator_data->pinned);
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
if (new_ptr == NULL)
{
#ifdef HAVE_SYNC_BUILTINS
__atomic_add_fetch (&allocator_data->used_pool_size,
prev_size - new_size,
MEMMODEL_RELAXED);
#else
gomp_mutex_lock (&allocator_data->lock);
allocator_data->used_pool_size -= new_size - prev_size;
gomp_mutex_unlock (&allocator_data->lock);
#endif
goto fail;
}
else if (prev_size)
{
ret = (char *) new_ptr + sizeof (struct omp_mem_header);
((struct omp_mem_header *) ret)[-1].ptr = new_ptr;
((struct omp_mem_header *) ret)[-1].size = new_size;
((struct omp_mem_header *) ret)[-1].allocator = allocator;
return ret;
}
}
else if (new_alignment == sizeof (void *)
&& old_alignment == sizeof (struct omp_mem_header)
#if defined(LIBGOMP_USE_MEMKIND) || defined(LIBGOMP_USE_LIBNUMA)
&& memkind == free_memkind
#endif
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
&& (free_allocator_data == NULL
|| free_allocator_data->pool_size == ~(uintptr_t) 0))
{
#ifdef LIBGOMP_USE_LIBNUMA
if (memkind == GOMP_MEMKIND_LIBNUMA)
new_ptr = libnuma_data->numa_realloc (data->ptr, data->size, new_size);
# ifdef LIBGOMP_USE_MEMKIND
else
# endif
#endif
#ifdef LIBGOMP_USE_MEMKIND
if (memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
void *kind = *memkind_data->kinds[memkind];
new_ptr = memkind_data->memkind_realloc (kind, data->ptr,
new_size);
}
else
#endif
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
{
omp_memspace_handle_t memspace;
memspace = (allocator_data
? allocator_data->memspace
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
: predefined_alloc_mapping (allocator));
int was_pinned = (free_allocator_data
? free_allocator_data->pinned
: free_allocator == ompx_gnu_pinned_mem_alloc);
int pinned = (allocator_data
? allocator_data->pinned
: allocator == ompx_gnu_pinned_mem_alloc);
new_ptr = MEMSPACE_REALLOC (memspace, data->ptr, data->size, new_size,
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
was_pinned, pinned);
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
}
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
if (new_ptr == NULL)
goto fail;
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
ret = (char *) new_ptr + sizeof (struct omp_mem_header);
((struct omp_mem_header *) ret)[-1].ptr = new_ptr;
((struct omp_mem_header *) ret)[-1].size = new_size;
((struct omp_mem_header *) ret)[-1].allocator = allocator;
return ret;
}
else
{
#ifdef LIBGOMP_USE_LIBNUMA
if (memkind == GOMP_MEMKIND_LIBNUMA)
new_ptr = libnuma_data->numa_alloc_local (new_size);
# ifdef LIBGOMP_USE_MEMKIND
else
# endif
#endif
#ifdef LIBGOMP_USE_MEMKIND
if (memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
void *kind = *memkind_data->kinds[memkind];
new_ptr = memkind_data->memkind_malloc (kind, new_size);
}
else
#endif
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
{
omp_memspace_handle_t memspace;
memspace = (allocator_data
? allocator_data->memspace
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
: predefined_alloc_mapping (allocator));
int pinned = (allocator_data
? allocator_data->pinned
: allocator == ompx_gnu_pinned_mem_alloc);
new_ptr = MEMSPACE_ALLOC (memspace, new_size, pinned);
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
}
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
if (new_ptr == NULL)
goto fail;
}
if (new_alignment > sizeof (void *))
ret = (void *) (((uintptr_t) new_ptr
+ sizeof (struct omp_mem_header)
+ new_alignment - sizeof (void *))
& ~(new_alignment - 1));
else
ret = (char *) new_ptr + sizeof (struct omp_mem_header);
((struct omp_mem_header *) ret)[-1].ptr = new_ptr;
((struct omp_mem_header *) ret)[-1].size = new_size;
((struct omp_mem_header *) ret)[-1].allocator = allocator;
if (old_size - old_alignment < size)
size = old_size - old_alignment;
memcpy (ret, ptr, size);
if (__builtin_expect (free_allocator_data
&& free_allocator_data->pool_size < ~(uintptr_t) 0, 0))
{
#ifdef HAVE_SYNC_BUILTINS
__atomic_add_fetch (&free_allocator_data->used_pool_size, -data->size,
MEMMODEL_RELAXED);
#else
gomp_mutex_lock (&free_allocator_data->lock);
free_allocator_data->used_pool_size -= data->size;
gomp_mutex_unlock (&free_allocator_data->lock);
#endif
}
#ifdef LIBGOMP_USE_LIBNUMA
if (free_memkind == GOMP_MEMKIND_LIBNUMA)
{
libnuma_data->numa_free (data->ptr, data->size);
return ret;
}
# ifdef LIBGOMP_USE_MEMKIND
else
# endif
#endif
#ifdef LIBGOMP_USE_MEMKIND
if (free_memkind)
{
struct gomp_memkind_data *memkind_data = gomp_get_memkind ();
void *kind = *memkind_data->kinds[free_memkind];
memkind_data->memkind_free (kind, data->ptr);
return ret;
}
#endif
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
{
omp_memspace_handle_t was_memspace;
was_memspace = (free_allocator_data
? free_allocator_data->memspace
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
: predefined_alloc_mapping (free_allocator));
int was_pinned = (free_allocator_data
? free_allocator_data->pinned
: free_allocator == ompx_gnu_pinned_mem_alloc);
MEMSPACE_FREE (was_memspace, data->ptr, data->size, was_pinned);
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
}
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
return ret;
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
fail:;
int fallback = (allocator_data
? allocator_data->fallback
libgomp, openmp: Add ompx_gnu_pinned_mem_alloc This creates a new predefined allocator as a shortcut for using pinned memory with OpenMP. This is not in the OpenMP standard so it uses the "ompx" namespace and an independent enum baseline of 200 (selected to not clash with other known implementations). The allocator is equivalent to using a custom allocator with the pinned trait and the null fallback trait. One motivation for having this feature is for use by the (planned) -foffload-memory=pinned feature. gcc/fortran/ChangeLog: * openmp.cc (is_predefined_allocator): Update valid ranges to incorporate ompx_gnu_pinned_mem_alloc. libgomp/ChangeLog: * allocator.c (ompx_gnu_min_predefined_alloc): New. (ompx_gnu_max_predefined_alloc): New. (predefined_alloc_mapping): Rename to ... (predefined_omp_alloc_mapping): ... this. (predefined_ompx_gnu_alloc_mapping): New. (_Static_assert): Adjust for the new name, and add a new assert for the new table. (predefined_allocator_p): New. (predefined_alloc_mapping): New. (omp_aligned_alloc): Support ompx_gnu_pinned_mem_alloc. Use predefined_allocator_p and predefined_alloc_mapping. (omp_free): Likewise. (omp_alligned_calloc): Likewise. (omp_realloc): Likewise. * env.c (parse_allocator): Add ompx_gnu_pinned_mem_alloc. * libgomp.texi: Document ompx_gnu_pinned_mem_alloc. * omp.h.in (omp_allocator_handle_t): Add ompx_gnu_pinned_mem_alloc. * omp_lib.f90.in: Add ompx_gnu_pinned_mem_alloc. * omp_lib.h.in: Add ompx_gnu_pinned_mem_alloc. * testsuite/libgomp.c/alloc-pinned-5.c: New test. * testsuite/libgomp.c/alloc-pinned-6.c: New test. * testsuite/libgomp.fortran/alloc-pinned-1.f90: New test. gcc/testsuite/ChangeLog: * gfortran.dg/gomp/allocate-pinned-1.f90: New test. Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2024-06-12 19:09:33 +08:00
: (allocator == omp_default_mem_alloc
|| allocator == ompx_gnu_pinned_mem_alloc)
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
? omp_atv_null_fb
: omp_atv_default_mem_fb);
switch (fallback)
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
{
libgomp, nvptx: low-latency memory allocator This patch adds support for allocating low-latency ".shared" memory on NVPTX GPU device, via the omp_low_lat_mem_space and omp_alloc. The memory can be allocated, reallocated, and freed using a basic but fast algorithm, is thread safe and the size of the low-latency heap can be configured using the GOMP_NVPTX_LOWLAT_POOL environment variable. The use of the PTX dynamic_smem_size feature means that low-latency allocator will not work with the PTX 3.1 multilib. For now, the omp_low_lat_mem_alloc allocator also works, but that will change when I implement the access traits. libgomp/ChangeLog: * allocator.c (MEMSPACE_ALLOC): New macro. (MEMSPACE_CALLOC): New macro. (MEMSPACE_REALLOC): New macro. (MEMSPACE_FREE): New macro. (predefined_alloc_mapping): New array. Add _Static_assert to match. (ARRAY_SIZE): New macro. (omp_aligned_alloc): Use MEMSPACE_ALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_free): Use MEMSPACE_FREE. (omp_calloc): Use MEMSPACE_CALLOC. Implement fall-backs for predefined allocators. Simplify existing fall-backs. (omp_realloc): Use MEMSPACE_REALLOC, MEMSPACE_ALLOC, and MEMSPACE_FREE. Implement fall-backs for predefined allocators. Simplify existing fall-backs. * config/nvptx/team.c (__nvptx_lowlat_pool): New asm variable. (__nvptx_lowlat_init): New prototype. (gomp_nvptx_main): Call __nvptx_lowlat_init. * libgomp.texi: Update memory space table. * plugin/plugin-nvptx.c (lowlat_pool_size): New variable. (GOMP_OFFLOAD_init_device): Read the GOMP_NVPTX_LOWLAT_POOL envvar. (GOMP_OFFLOAD_run): Apply lowlat_pool_size. * basic-allocator.c: New file. * config/nvptx/allocator.c: New file. * testsuite/libgomp.c/omp_alloc-1.c: New test. * testsuite/libgomp.c/omp_alloc-2.c: New test. * testsuite/libgomp.c/omp_alloc-3.c: New test. * testsuite/libgomp.c/omp_alloc-4.c: New test. * testsuite/libgomp.c/omp_alloc-5.c: New test. * testsuite/libgomp.c/omp_alloc-6.c: New test. Co-authored-by: Kwok Cheung Yeung <kcy@codesourcery.com> Co-Authored-By: Thomas Schwinge <thomas@codesourcery.com>
2021-12-04 01:46:41 +08:00
case omp_atv_default_mem_fb:
allocator = omp_default_mem_alloc;
goto retry;
case omp_atv_null_fb:
break;
default:
case omp_atv_abort_fb:
gomp_fatal ("Out of memory allocating %lu bytes",
(unsigned long) size);
case omp_atv_allocator_fb:
allocator = allocator_data->fb_data;
goto retry;
openmp: Add omp_aligned_{,c}alloc and omp_{c,re}alloc This patch adds new OpenMP 5.1 allocator entrypoints and in addition to that fixes an omp_alloc bug which is hard to test for - if the first allocator fails but has a larger alignment trait and has a fallback allocator, either the default behavior or a user fallback, then the extra alignment will be used even in the fallback allocation, rather than just starting with whatever alignment has been requested (in GOMP_alloc or the minimum one in omp_alloc). Jonathan's comment on IRC this morning made me realize that I should add alloc_align attributes to 2 of the prototypes and I still need to add testsuite coverage for omp_realloc, will do that in a follow-up. 2021-09-30 Jakub Jelinek <jakub@redhat.com> * omp.h.in (omp_aligned_alloc, omp_calloc, omp_aligned_calloc, omp_realloc): New prototypes. (omp_alloc): Move after omp_free prototype, add __malloc__ (omp_free) attribute. * allocator.c: Include string.h. (omp_aligned_alloc): No longer static, add ialias. Add new_alignment variable and use it instead of alignment so that when retrying the old alignment is used again. Don't retry if new alignment is the same as old alignment, unless allocator had pool size. (omp_alloc, GOMP_alloc, GOMP_free): Use ialias_call. (omp_aligned_calloc, omp_calloc, omp_realloc): New functions. * libgomp.map (OMP_5.0.2): Export omp_aligned_alloc, omp_calloc, omp_aligned_calloc and omp_realloc. * testsuite/libgomp.c-c++-common/alloc-4.c (main): Add omp_aligned_alloc, omp_calloc and omp_aligned_calloc tests. * testsuite/libgomp.c-c++-common/alloc-5.c: New test. * testsuite/libgomp.c-c++-common/alloc-6.c: New test. * testsuite/libgomp.c-c++-common/alloc-7.c: New test. * testsuite/libgomp.c-c++-common/alloc-8.c: New test.
2021-09-30 15:30:18 +08:00
}
return NULL;
openmp: Implement allocate clause in omp lowering. For now, task/taskloop constructs aren't handled and C/C++ array reductions and reductions with task or inscan modifiers need further work. Instead of calling omp_alloc/omp_free (where the former doesn't have alignment argument and omp_aligned_alloc is 5.1 only feature), this calls GOMP_alloc/GOMP_free, so that the library can fail if it would fall back into NULL (exception is zero length allocations). 2020-11-12 Jakub Jelinek <jakub@redhat.com> gcc/ * builtin-types.def (BT_FN_PTR_SIZE_SIZE_PTRMODE): New function type. * omp-builtins.def (BUILT_IN_GOACC_DECLARE): Move earlier. (BUILT_IN_GOMP_ALLOC, BUILT_IN_GOMP_FREE): New builtins. * gimplify.c (gimplify_scan_omp_clauses): Force allocator into a decl if it is not NULL, INTEGER_CST or decl. (gimplify_adjust_omp_clauses): Clear GOVD_EXPLICIT on explicit clauses which are being removed. Remove allocate clauses for variables not seen if they are private, firstprivate or linear too. Call omp_notice_variable on the allocator otherwise. (gimplify_omp_for): Handle iterator vars mentioned in allocate clauses similarly to non-is_gimple_reg iterators. * omp-low.c (struct omp_context): Add allocate_map field. (delete_omp_context): Delete it. (scan_sharing_clauses): Fill it from allocate clauses. Remove it if mentioned also in shared clause. (lower_private_allocate): New function. (lower_rec_input_clauses): Handle allocate clause for privatized variables, except for task/taskloop, C/C++ array reductions for now and task/inscan variables. (lower_send_shared_vars): Don't consider variables in allocate_map as shared. * omp-expand.c (expand_omp_for_generic, expand_omp_for_static_nochunk, expand_omp_for_static_chunk): Use expand_omp_build_assign instead of gimple_build_assign + gsi_insert_after. * builtins.c (builtin_fnspec): Handle BUILTIN_GOMP_ALLOC and BUILTIN_GOMP_FREE. * tree-ssa-ccp.c (evaluate_stmt): Handle BUILTIN_GOMP_ALLOC. * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Handle BUILTIN_GOMP_ALLOC. (mark_all_reaching_defs_necessary_1): Handle BUILTIN_GOMP_ALLOC and BUILTIN_GOMP_FREE. (propagate_necessity): Likewise. gcc/fortran/ * f95-lang.c (ATTR_ALLOC_WARN_UNUSED_RESULT_SIZE_2_NOTHROW_LIST): Define. (gfc_init_builtin_functions): Add alloc_size and warn_unused_result attributes to __builtin_GOMP_alloc. * types.def (BT_PTRMODE): New primitive type. (BT_FN_VOID_PTR_PTRMODE, BT_FN_PTR_SIZE_SIZE_PTRMODE): New function types. libgomp/ * libgomp.map (GOMP_alloc, GOMP_free): Export at GOMP_5.0.1. * omp.h.in (omp_alloc): Add malloc and alloc_size attributes. * libgomp_g.h (GOMP_alloc, GOMP_free): Declare. * allocator.c (omp_aligned_alloc): New for now static function, add alignment argument and handle it. (omp_alloc): Reimplement using omp_aligned_alloc. (GOMP_alloc, GOMP_free): New functions. (omp_free): Add ialias. * testsuite/libgomp.c-c++-common/allocate-1.c: New test. * testsuite/libgomp.c++/allocate-1.C: New test.
2020-11-13 04:38:04 +08:00
}