Chung-Lin Tang [Fri, 17 Jun 2022 14:22:25 +0000 (22:22 +0800)]
openmp: Implement uses_allocators clause
This is a merge of:
https://gcc.gnu.org/pipermail/gcc-patches/2022-June/596412.html
For user defined allocator handles, this allows target regions to assign
memory space and traits to allocators, and automatically calls
omp_init/destroy_allocator() in the beginning/end of the target region.
For pre-defined allocators (i.e. omp_..._mem_alloc names), this is a no-op,
such clauses are not created.
Asides from the front-end portions, the target region transforms are
done in gimplify_omp_workshare.
This patch also includes added changes to enforce the "allocate allocator
must be also in a uses_allocator clause". This is done during
gimplify_scan_omp_clauses.
* c-parser.cc (c_parser_omp_clause_name): Add case for uses_allocators
clause.
(c_parser_omp_clause_uses_allocators): New function.
(c_parser_omp_all_clauses): Add PRAGMA_OMP_CLAUSE_USES_ALLOCATORS case.
(OMP_TARGET_CLAUSE_MASK): Add PRAGMA_OMP_CLAUSE_USES_ALLOCATORS to mask.
* c-typeck.cc (c_finish_omp_clauses): Add case handling for
OMP_CLAUSE_USES_ALLOCATORS.
gcc/cp/ChangeLog:
* parser.cc (cp_parser_omp_clause_name): Add case for uses_allocators
clause.
(cp_parser_omp_clause_uses_allocators): New function.
(cp_parser_omp_all_clauses): Add PRAGMA_OMP_CLAUSE_USES_ALLOCATORS case.
(OMP_TARGET_CLAUSE_MASK): Add PRAGMA_OMP_CLAUSE_USES_ALLOCATORS to mask.
* semantics.cc (finish_omp_clauses): Add case handling for
OMP_CLAUSE_USES_ALLOCATORS.
fortran/ChangeLog:
* gfortran.h (struct gfc_omp_namelist): Add memspace_sym, traits_sym
fields.
(OMP_LIST_USES_ALLOCATORS): New list enum.
* openmp.cc (enum omp_mask2): Add OMP_CLAUSE_USES_ALLOCATORS.
(gfc_match_omp_clause_uses_allocators): New function.
(gfc_match_omp_clauses): Add case to handle OMP_CLAUSE_USES_ALLOCATORS.
(OMP_TARGET_CLAUSES): Add OMP_CLAUSE_USES_ALLOCATORS.
(resolve_omp_clauses): Add "USES_ALLOCATORS" to clause_names[].
* dump-parse-tree.cc (show_omp_namelist): Handle OMP_LIST_USES_ALLOCATORS.
(show_omp_clauses): Likewise.
* trans-array.cc (gfc_conv_array_initializer): Adjust array index
to always be a created tree expression instead of NULL_TREE when zero.
* trans-openmp.cc (gfc_trans_omp_clauses): For ALLOCATE clause, handle
using gfc_trans_omp_variable for EXPR_VARIABLE exprs.
Add handling of OMP_LIST_USES_ALLOCATORS case.
* types.def (BT_FN_VOID_PTRMODE): Define.
(BT_FN_PTRMODE_PTRMODE_INT_PTR): Define.
* gimplify.cc (gimplify_scan_omp_clauses): Add checking of OpenMP target
region allocate clauses, to require a uses_allocators clause to exist
for allocators.
(gimplify_omp_workshare): Add handling of OMP_CLAUSE_USES_ALLOCATORS
for OpenMP target regions; create calls of omp_init/destroy_allocator
around target region body.
* omp-low.cc (lower_private_allocate): Adjust receiving of allocator.
(lower_rec_input_clauses): Likewise.
(create_task_copyfn): Add dereference for allocator if needed.
* system.h (startswith): New function.
gcc/testsuite/ChangeLog:
* c-c++-common/gomp/uses_allocators-1.c: New test.
* c-c++-common/gomp/uses_allocators-2.c: New test.
* c-c++-common/gomp/uses_allocators-3.c: New test.
* gfortran.dg/gomp/allocate-1.f90: Adjust testcase.
* gfortran.dg/gomp/uses_allocators-1.f90: New test.
* gfortran.dg/gomp/uses_allocators-2.f90: New test.
* gfortran.dg/gomp/uses_allocators-3.f90: New test.
Chung-Lin Tang [Fri, 17 Jun 2022 14:17:47 +0000 (22:17 +0800)]
middle-end/103851 - ensure SSA names are released during OMP lowering
This makes sure to release moved & remapped SSA names during OMP
outlining which happens before going into SSA but with SSA names
created by gimplification around.
Chung-Lin Tang [Mon, 13 Jun 2022 13:12:32 +0000 (21:12 +0800)]
openmp: Implement uses_allocators clause
This is a merge of:
https://gcc.gnu.org/pipermail/gcc-patches/2022-June/596412.html
For user defined allocator handles, this allows target regions to assign
memory space and traits to allocators, and automatically calls
omp_init/destroy_allocator() in the beginning/end of the target region.
For pre-defined allocators (i.e. omp_..._mem_alloc names), this is a no-op,
such clauses are not created.
Asides from the front-end portions, the target region transforms are
done in gimplify_omp_workshare.
This patch also includes added changes to enforce the "allocate allocator
must be also in a uses_allocator clause". This is done during
gimplify_scan_omp_clauses.
* c-parser.cc (c_parser_omp_clause_name): Add case for uses_allocators
clause.
(c_parser_omp_clause_uses_allocators): New function.
(c_parser_omp_all_clauses): Add PRAGMA_OMP_CLAUSE_USES_ALLOCATORS case.
(OMP_TARGET_CLAUSE_MASK): Add PRAGMA_OMP_CLAUSE_USES_ALLOCATORS to mask.
* c-typeck.cc (c_finish_omp_clauses): Add case handling for
OMP_CLAUSE_USES_ALLOCATORS.
gcc/cp/ChangeLog:
* parser.cc (cp_parser_omp_clause_name): Add case for uses_allocators
clause.
(cp_parser_omp_clause_uses_allocators): New function.
(cp_parser_omp_all_clauses): Add PRAGMA_OMP_CLAUSE_USES_ALLOCATORS case.
(OMP_TARGET_CLAUSE_MASK): Add PRAGMA_OMP_CLAUSE_USES_ALLOCATORS to mask.
* semantics.cc (finish_omp_clauses): Add case handling for
OMP_CLAUSE_USES_ALLOCATORS.
fortran/ChangeLog:
* gfortran.h (struct gfc_omp_namelist): Add memspace_sym, traits_sym
fields.
(OMP_LIST_USES_ALLOCATORS): New list enum.
* openmp.cc (enum omp_mask2): Add OMP_CLAUSE_USES_ALLOCATORS.
(gfc_match_omp_clause_uses_allocators): New function.
(gfc_match_omp_clauses): Add case to handle OMP_CLAUSE_USES_ALLOCATORS.
(OMP_TARGET_CLAUSES): Add OMP_CLAUSE_USES_ALLOCATORS.
(resolve_omp_clauses): Add "USES_ALLOCATORS" to clause_names[].
* dump-parse-tree.cc (show_omp_namelist): Handle OMP_LIST_USES_ALLOCATORS.
(show_omp_clauses): Likewise.
* trans-array.cc (gfc_conv_array_initializer): Adjust array index
to always be a created tree expression instead of NULL_TREE when zero.
* trans-openmp.cc (gfc_trans_omp_clauses): For ALLOCATE clause, handle
using gfc_trans_omp_variable for EXPR_VARIABLE exprs.
Add handling of OMP_LIST_USES_ALLOCATORS case.
* types.def (BT_FN_VOID_PTRMODE): Define.
(BT_FN_PTRMODE_PTRMODE_INT_PTR): Define.
* gimplify.cc (gimplify_scan_omp_clauses): Add checking of OpenMP target
region allocate clauses, to require a uses_allocators clause to exist
for allocators.
(gimplify_omp_workshare): Add handling of OMP_CLAUSE_USES_ALLOCATORS
for OpenMP target regions; create calls of omp_init/destroy_allocator
around target region body.
* omp-low.cc (lower_private_allocate): Adjust receiving of allocator.
(lower_rec_input_clauses): Likewise.
(create_task_copyfn): Add dereference for allocator if needed.
* system.h (startswith): New function.
gcc/testsuite/ChangeLog:
* c-c++-common/gomp/uses_allocators-1.c: New test.
* c-c++-common/gomp/uses_allocators-2.c: New test.
* c-c++-common/gomp/uses_allocators-3.c: New test.
* gfortran.dg/gomp/allocate-1.f90: Adjust testcase.
* gfortran.dg/gomp/uses_allocators-1.f90: New test.
* gfortran.dg/gomp/uses_allocators-2.f90: New test.
* gfortran.dg/gomp/uses_allocators-3.f90: New test.
Andrew Jenner [Mon, 30 May 2022 10:12:15 +0000 (10:12 +0000)]
[PATCH] GCN: Implement __atomic_compare_exchange_{1,2} in libgcc [PR102215]
libgcc/ChangeLog:
PR target/102215
* config/gcn/atomic.c (__sync_val_compare_and_swap_##SIZE): Move
a line up to non-arg-dependent value first.
(__ATOMIC_COMPARE_EXCHANGE): Define + call to generate
__atomic_compare_exchange_{1,2}.
Andrew Stubbs [Thu, 24 Feb 2022 17:16:13 +0000 (17:16 +0000)]
amdgcn: Add gfx90a support
This adds architecture options and multilibs for the AMD GFX90a GPUs.
It also tidies up some of the ISA selection code, and corrects a few small
mistake in the gfx908 naming.
gcc/ChangeLog:
* config.gcc (amdgcn): Accept --with-arch=gfx908 and gfx90a.
* config/gcn/gcn-opts.h (enum gcn_isa): New.
(TARGET_GCN3): Use enum gcn_isa.
(TARGET_GCN3_PLUS): Likewise.
(TARGET_GCN5): Likewise.
(TARGET_GCN5_PLUS): Likewise.
(TARGET_CDNA1): New.
(TARGET_CDNA1_PLUS): New.
(TARGET_CDNA2): New.
(TARGET_CDNA2_PLUS): New.
(TARGET_M0_LDS_LIMIT): New.
(TARGET_PACKED_WORK_ITEMS): New.
* config/gcn/gcn.cc (gcn_isa): Change to enum gcn_isa.
(gcn_option_override): Recognise CDNA ISA variants.
(gcn_omp_device_kind_arch_isa): Support gfx90a.
(gcn_expand_prologue): Make m0 init optional.
Add support for packed work items.
(output_file_start): Support gfx90a.
(gcn_hsa_declare_function_name): Support gfx90a metadata.
* config/gcn/gcn.h (TARGET_CPU_CPP_BUILTINS):Add __CDNA1__ and
__CDNA2__.
* config/gcn/gcn.md (<su>mulsi3_highpart): Use TARGET_GCN5_PLUS.
(<su>mulsi3_highpart_imm): Likewise.
(<su>mulsidi3): Likewise.
(<su>mulsidi3_imm): Likewise.
* config/gcn/gcn.opt (gpu_type): Add gfx90a.
* config/gcn/mkoffload.cc (EF_AMDGPU_MACH_AMDGCN_GFX90a): New.
(main): Support gfx90a.
* config/gcn/t-gcn-hsa: Add gfx90a multilib.
* config/gcn/t-omp-device: Add gfx90a isa.
Tobias Burnus [Thu, 12 May 2022 12:49:29 +0000 (14:49 +0200)]
OpenMP: Handle descriptors in target's firstprivate [PR104949]
For allocatable/pointer arrays, a firstprivate to a device
not only needs to privatize the descriptor but also the actual
data. This is implemented as:
firstprivate(x) firstprivate(x.data) attach(x [bias: &x.data-&x)
where the address of x in device memory is saved in hostaddrs[i]
by libgomp and the middle end actually passes hostaddrs[i]' to
attach.
OG11 version of the GCC 13/mainline patch submitted at
https://gcc.gnu.org/pipermail/gcc-patches/2022-May/594582.html
Note: Contrary to GCC 12+, OG11 does not support has_device_addr,
which is the main difference to the posted patch.
PR fortran/104949
* target.c (gomp_map_vars_internal, copy_firstprivate_data):
Support attach for GOMP_MAP_FIRSTPRIVATE.
* testsuite/libgomp.fortran/target-firstprivate-1.f90: New test.
* testsuite/libgomp.fortran/target-firstprivate-2.f90: New test.
* testsuite/libgomp.fortran/target-firstprivate-3.f90: New test.
Tobias Burnus [Thu, 12 May 2022 08:39:58 +0000 (10:39 +0200)]
Fortran: Fix proc pointer as elemental arg handling
The vtab's _callback function calls the elemental 'cb'
cb (var(:)%comp, comp_types_vtable._callback);
which gets called in a scalarization loop as 'var' might be a
nonscalar. Without the patch, that got translated as:
D.1234 = &comp_types_vtable._callback
...
cb (&(*D.4060)[S.3 + D.4071], &D.1234);
where 'D.1234' is function_type. With the patch, it remains a pointer;
i.e. D.1234 = comp... and 'cb (..., D.1234)', avoiding ME ICE.
Note: Fortran (F2018, C15100) requires that dummy arguments are
dummy data objects, which rules out dummy procs/proc-pointer dummies,
which is enforced in resolve_fl_procedure.
Thus, this change only affects the internally generated code.
gcc/fortran/ChangeLog:
* trans-array.c (gfc_scalar_elemental_arg_saved_as_reference):
Return true for attr.proc_pointer expressions.
gcc/testsuite/ChangeLog:
* gfortran.dg/finalize_38.f90: Compile with -Ofast.
* gfortran.dg/abstract_type_6.f03: Remove dg-error as
now hidden by other errors; copy to ...
* gfortran.dg/abstract_type_6a.f03: ... here; remove
some error to diagnose the error.
* gfortran.dg/finalize_39.f90: New test.
Fortran: Fix finalization resolution with deep copy
Follow-up patch to
"Fortran/OpenMP: Support mapping of DT with allocatable components"
https://gcc.gnu.org/pipermail/gcc-patches/2022-March/591144.html
gcc/fortran/ChangeLog:
* resolve.c (gfc_resolve_finalizers): Also resolve allocatable comps.
OpenMP: Fix use_device_{addr,ptr} with in-data-sharing arg
For array-descriptor vars, the descriptor is assigned to a temporary. However,
this failed when the clause's argument was in turn in a data-sharing clause
as the outer context's VALUE_EXPR wasn't used.
Andrew Stubbs [Wed, 13 Apr 2022 15:55:47 +0000 (16:55 +0100)]
openmp: unified_address support
This makes "requires unified_address" work by making it eqivalent to
"requires unified_shared_memory". This is more than is strictly necessary,
but should be standard compliant.
gcc/c/ChangeLog:
* c-parser.c (c_parser_omp_requires): Check requires unified_address
for conflict with -foffload-memory=shared.
gcc/cp/ChangeLog:
* parser.c (cp_parser_omp_requires): Check requires unified_address
for conflict with -foffload-memory=shared.
gcc/fortran/ChangeLog:
* openmp.c (gfc_match_omp_requires): Check requires unified_address
for conflict with -foffload-memory=shared.
gcc/ChangeLog:
* omp-low.c: Do USM transformations for "unified_address".
gcc/testsuite/ChangeLog:
* c-c++-common/gomp/usm-4.c: New test.
* gfortran.dg/gomp/usm-4.f90: New test.
This patch fixes a bug in lower_omp_target, where for Fortran arrays,
the expanded sender assignment is wrongly using the variable in the
current ctx, instead of the one looked-up outside, which is causing
use_device_ptr/addr to fail to work when used inside an omp-parallel
(where the omp child_fn is split away from the original).
The fix is inside omp-low.cc, though because the omp_array_data langhook
is used only by Fortran, this is essentially Fortran-specific.
* omp-low.cc (lower_omp_target): Use outer context looked-up 'var' as
argument to lang_hooks.decls.omp_array_data, instead of 'ovar' from
current clause.
libgomp/ChangeLog:
* testsuite/libgomp.fortran/use_device_ptr-4.f90: New testcase.
Fix a crash due to mismatch of free and GOMP_alloc.
With allocate directive, we replace the malloc calls to GOMP_alloc if
it is associated with the allocate statement. The memory was supposed
to be free-d by the implicitely generated free calls which also get
replaced. But if user explicitely deallocated the memory using the
deallocate statement, it can cause a mismatch. This commit handles
that case and also replaces the free call generated for deallocate
clause.
Also added deallocate in the testcase and tidied it up a bit.
gcc/ChangeLog.omp:
* omp-low.c (lower_omp_allocate): Move allocate declaration
inside loop. Set it to false at the end of condition.
Tobias Burnus [Wed, 23 Mar 2022 08:44:39 +0000 (09:44 +0100)]
LTO: Fixes for renaming issues with offload/OpenMP [PR104285]
gcc/lto/ChangeLog:
PR middle-end/104285
* lto-partition.c (maybe_rewrite_identifier): Use get_identifier
for the returned string to be usable as hash key.
(validize_symbol_for_target): Hence, use return value directly.
(privatize_symbol_name_1): Track maybe_rewrite_identifier renames.
* lto.c (offload_handle_link_vars): Move function up before ...
(do_whole_program_analysis): Call it after static renamings.
(lto_main): Move call after static renamings.
libgomp/ChangeLog:
PR middle-end/104285
* testsuite/libgomp.c++/target-same-name-2-a.C: New test.
* testsuite/libgomp.c++/target-same-name-2-b.C: New test.
* testsuite/libgomp.c++/target-same-name-2.C: New test.
* testsuite/libgomp.c-c++-common/target-same-name-1-a.c: New test.
* testsuite/libgomp.c-c++-common/target-same-name-1-b.c: New test.
* testsuite/libgomp.c-c++-common/target-same-name-1.c: New test.
Andrew Stubbs [Fri, 11 Mar 2022 12:58:38 +0000 (12:58 +0000)]
openmp: -foffload-memory=pinned
Implement the -foffload-memory=pinned option such that libgomp is
instructed to enable fully-pinned memory at start-up. The option is
intended to provide a performance boost to certain offload programs without
modifying the code.
This feature only works on Linux, at present, and simply calls mlockall to
enable always-on memory pinning. It requires that the ulimit feature is
set high enough to accommodate all the program's memory usage.
In this mode the ompx_pinned_memory_alloc feature is disabled as it is not
needed and may conflict.
Backport of the patch posted at
https://gcc.gnu.org/pipermail/gcc-patches/2022-March/591354.html
gcc/ChangeLog:
* omp-low.c (omp_enable_pinned_mode): New function.
(execute_lower_omp): Call omp_enable_pinned_mode.
libgomp/ChangeLog:
* config/linux/allocator.c (always_pinned_mode): New variable.
(GOMP_enable_pinned_mode): New function.
(linux_memspace_alloc): Disable pinning when always_pinned_mode set.
(linux_memspace_calloc): Likewise.
(linux_memspace_free): Likewise.
(linux_memspace_realloc): Likewise.
* libgomp.map (GOMP_5.1.1): New version space with
GOMP_enable_pinned_mode.
* testsuite/libgomp.c/alloc-pinned-7.c: New test.
openmp: Use libgomp memory allocation functions with unified shared memory.
This patches changes calls to malloc/free/calloc/realloc/aligned_alloc and
operator new to memory allocation functions in libgomp with
allocator=ompx_unified_shared_mem_alloc. This helps existing code to benefit
from the unified shared memory. The libgomp does the correct thing with all
the mapping constructs and there is no memory copies if the pointer is pointing
to unified shared memory.
We only replace replacable new operator and not the class member or placement new.
Backport of a patch posted at
https://gcc.gnu.org/pipermail/gcc-patches/2022-March/591353.html
gcc/ChangeLog:
* omp-low.c (usm_transform): New function.
(make_pass_usm_transform): Likewise.
(class pass_usm_transform): New.
* passes.def: Add pass_usm_transform.
* tree-pass.h (make_pass_usm_transform): New declaration.
gcc/testsuite/ChangeLog:
* c-c++-common/gomp/usm-2.c: New test.
* c-c++-common/gomp/usm-3.c: New test.
* g++.dg/gomp/usm-1.C: New test.
* g++.dg/gomp/usm-2.C: New test.
* g++.dg/gomp/usm-3.C: New test.
* gfortran.dg/gomp/usm-2.f90: New test.
* gfortran.dg/gomp/usm-3.f90: New test.
libgomp/ChangeLog:
* testsuite/libgomp.c/usm-6.c: New test.
* testsuite/libgomp.c++/usm-1.C: Likewise.
Andrew Stubbs [Fri, 11 Mar 2022 12:37:58 +0000 (12:37 +0000)]
openmp, nvptx: ompx_unified_shared_mem_alloc
This adds support for using Cuda Managed Memory with omp_alloc. It will be
used as the underpinnings for "requires unified_shared_memory" in a later
patch.
There are two new predefined allocators, ompx_unified_shared_mem_alloc and
ompx_host_mem_alloc, plus corresponding memory spaces, which can be used to
allocate memory in the "managed" space and explicitly on the host (it is
intended that "malloc" will be intercepted by the compiler).
The nvptx plugin is modified to make the necessary Cuda calls, and libgomp
is modified to switch to shared-memory mode for USM allocated mappings.
Backport of the patch posted at
https://gcc.gnu.org/pipermail/gcc-patches/2022-March/591352.html
libgomp/ChangeLog:
* allocator.c (omp_max_predefined_alloc): Update.
(omp_aligned_alloc): Don't fallback ompx_host_mem_alloc.
(omp_aligned_calloc): Likewise.
(omp_realloc): Likewise.
* config/linux/allocator.c (linux_memspace_alloc): Handle USM.
(linux_memspace_calloc): Handle USM.
(linux_memspace_free): Handle USM.
(linux_memspace_realloc): Handle USM.
* config/nvptx/allocator.c (nvptx_memspace_alloc): Reject
ompx_host_mem_alloc.
(nvptx_memspace_calloc): Likewise.
(nvptx_memspace_realloc): Likewise.
* libgomp-plugin.h (GOMP_OFFLOAD_usm_alloc): New prototype.
(GOMP_OFFLOAD_usm_free): New prototype.
(GOMP_OFFLOAD_is_usm_ptr): New prototype.
* libgomp.h (gomp_usm_alloc): New prototype.
(gomp_usm_free): New prototype.
(gomp_is_usm_ptr): New prototype.
(struct gomp_device_descr): Add USM functions.
* omp.h.in (omp_memspace_handle_t): Add ompx_unified_shared_mem_space
and ompx_host_mem_space.
(omp_allocator_handle_t): Add ompx_unified_shared_mem_alloc and
ompx_host_mem_alloc.
* omp_lib.f90.in: Likewise.
* plugin/plugin-nvptx.c (nvptx_alloc): Add "usm" parameter.
Call cuMemAllocManaged as appropriate.
(GOMP_OFFLOAD_alloc): Move internals to ...
(GOMP_OFFLOAD_alloc_1): ... this, and add usm parameter.
(GOMP_OFFLOAD_usm_alloc): New function.
(GOMP_OFFLOAD_usm_free): New function.
(GOMP_OFFLOAD_is_usm_ptr): New function.
* target.c (gomp_map_vars_internal): Add USM support.
(gomp_usm_alloc): New function.
(gomp_usm_free): New function.
(gomp_load_plugin_for_device): New function.
* testsuite/libgomp.c/usm-1.c: New test.
* testsuite/libgomp.c/usm-2.c: New test.
* testsuite/libgomp.c/usm-3.c: New test.
* testsuite/libgomp.c/usm-4.c: New test.
* testsuite/libgomp.c/usm-5.c: New test.
Andrew Stubbs [Fri, 11 Mar 2022 14:33:11 +0000 (14:33 +0000)]
openmp, nvptx: low-lat memory access traits
The NVPTX low latency memory is not accessible outside the team that allocates
it, and therefore should be unavailable for allocators with the access trait
"all". This change means that the omp_low_lat_mem_alloc predefined
allocator now implicitly implies the "pteam" trait.
Backport of a patch posted at
https://gcc.gnu.org/pipermail/gcc-patches/2022-January/589355.html
libgomp/ChangeLog:
* allocator.c (MEMSPACE_VALIDATE): New macro.
(omp_aligned_alloc): Use MEMSPACE_VALIDATE.
(omp_aligned_calloc): Likewise.
(omp_realloc): Likewise.
* config/nvptx/allocator.c (nvptx_memspace_validate): New function.
(MEMSPACE_VALIDATE): New macro.
* testsuite/libgomp.c/allocators-4.c (main): Add access trait.
* testsuite/libgomp.c/allocators-6.c (main): Add access trait.
* testsuite/libgomp.c/allocators-7.c: New test.
Andrew Stubbs [Fri, 11 Mar 2022 12:33:06 +0000 (12:33 +0000)]
libgomp, openmp: Add ompx_pinned_mem_alloc
This creates a new predefined allocator as a shortcut for using pinned
memory with OpenMP. The name uses the OpenMP extension space and is
intended to be consistent with other OpenMP implementations currently in
development.
The allocator is equivalent to using a custom allocator with the pinned
trait and the null fallback trait.
Backport of a patch posted at
https://gcc.gnu.org/pipermail/gcc-patches/2022-January/588951.html
* allocator.c (omp_max_predefined_alloc): Update.
(omp_aligned_alloc): Support ompx_pinned_mem_alloc.
(omp_free): Likewise.
(omp_aligned_calloc): Likewise.
(omp_realloc): Likewise.
* omp.h.in (omp_allocator_handle_t): Add ompx_pinned_mem_alloc.
* omp_lib.f90.in: Add ompx_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.
Andrew Stubbs [Fri, 11 Mar 2022 12:12:39 +0000 (12:12 +0000)]
libgomp: pinned memory
Implement the OpenMP pinned memory trait on Linux hosts using the mlock
syscall. Pinned allocations are performed using mmap, not malloc, to ensure
that they can be unpinned safely when freed.
Backport of a patch posted at
https://gcc.gnu.org/pipermail/gcc-patches/2022-January/588360.html
libgomp/ChangeLog:
* allocator.c (MEMSPACE_ALLOC): Add PIN.
(MEMSPACE_CALLOC): Add PIN.
(MEMSPACE_REALLOC): Add PIN.
(MEMSPACE_FREE): Add PIN.
(xmlock): New function.
(omp_init_allocator): Don't disallow the pinned trait.
(omp_aligned_alloc): Add pinning to all MEMSPACE_* calls.
(omp_aligned_calloc): Likewise.
(omp_realloc): Likewise.
(omp_free): Likewise.
* config/linux/allocator.c: New file.
* config/nvptx/allocator.c (MEMSPACE_ALLOC): Add PIN.
(MEMSPACE_CALLOC): Add PIN.
(MEMSPACE_REALLOC): Add PIN.
(MEMSPACE_FREE): Add PIN.
* testsuite/libgomp.c/alloc-pinned-1.c: New test.
* testsuite/libgomp.c/alloc-pinned-2.c: New test.
* testsuite/libgomp.c/alloc-pinned-3.c: New test.
* testsuite/libgomp.c/alloc-pinned-4.c: New test.
Chung-Lin Tang [Fri, 11 Mar 2022 13:50:52 +0000 (21:50 +0800)]
openmp: Allow classes with static members to be mappable
In OpenMP 5.x, static members are supposed to be not a barrier for a class
to be target-mapped. Remove the check for static members inside
cp_omp_mappable_type_1, and adjusts a testcase.
Add case for OMP_CLAUSE_ALLOCATOR in walk_tree_1. This helps fix
an ICE which occurs only on OG11 with allocate directive.
Please note that this change is not needed on master. The code
there handles all clauses in the same way so a special case for
OMP_CLAUSE_ALLOCATOR is not required.
gcc/
* tree.c (walk_tree_1): Add case for OMP_CLAUSE_ALLOCATOR.
Backport of a patch posted at
https://gcc.gnu.org/pipermail/gcc-patches/2022-January/588372.html
This patch looks for malloc/free calls that were generated by allocate statement
that is associated with allocate directive and replaces them with GOMP_alloc
and GOMP_free.
gcc/ChangeLog:
* omp-low.c (scan_sharing_clauses): Handle OMP_CLAUSE_ALLOCATOR.
(scan_omp_allocate): New.
(scan_omp_1_stmt): Call it.
(lower_omp_allocate): New function.
(lower_omp_1): Call it.
gcc/testsuite/ChangeLog:
* gfortran.dg/gomp/allocate-6.f90: Add tests.
libgomp/ChangeLog:
* testsuite/libgomp.fortran/allocate-2.f90: New test.
* gfortran.dg/gomp/allocate-7.f90: New test.
* gfortran.dg/gomp/allocate-8.f90: New test.
Handle cleanup of omp allocated variables (OpenMP 5.0).
Currently we are only handling omp allocate directive that is associated
with an allocate statement. This statement results in malloc and free calls.
The malloc calls are easy to get to as they are in the same block as allocate
directive. But the free calls come in a separate cleanup block. To help any
later passes finding them, an allocate directive is generated in the
cleanup block with kind=free. The normal allocate directive is given
kind=allocate.
Backport of a patch posted at
https://gcc.gnu.org/pipermail/gcc-patches/2022-January/588370.html
gcc/fortran/ChangeLog:
* gfortran.h (struct access_ref): Declare new members
omp_allocated and omp_allocated_end.
* openmp.c (gfc_match_omp_allocate): Set new_st.resolved_sym to
NULL.
(prepare_omp_allocated_var_list_for_cleanup): New function.
(gfc_resolve_omp_allocate): Call it.
* trans-decl.c (gfc_trans_deferred_vars): Process omp_allocated.
* trans-openmp.c (gfc_trans_omp_allocate): Set kind for the stmt
generated for allocate directive.
This is backport of a patch posted in
https://gcc.gnu.org/pipermail/gcc-patches/2022-February/590597.html
An allocate clause in target region must specify an allocator
unless the compilation unit has requires construct with
dynamic_allocators clause. Current implementation of the allocate
clause did not check for this restriction. This patch fills that
gap.
gcc/ChangeLog:
* omp-low.c (omp_maybe_offloaded_ctx): New prototype.
(scan_sharing_clauses): Check a restriction on allocate clause.
Fix multiple issue in the testcase allocate-1.f90.
This is backport of patch posted in
https://gcc.gnu.org/pipermail/gcc-patches/2022-February/589928.html
1. Thomas reported in
https://gcc.gnu.org/pipermail/gcc-patches/2022-January/589039.html
that this testcase is randomly failing. The problem was fixed pool
size which was exhausted when there were a lot of threads. Fixed it
by removing pool_size trait which causes default pool size to be used
which should be big enough.
2. Array indices have been changed to check the last element in the
array.
3. Remove a redundant assignment and move some code to better match
C testcase.
libgomp/ChangeLog:
* testsuite/libgomp.fortran/allocate-1.f90: Remove pool_size
trait. Test last index in w and v array. Remove redundant
assignment to V(1). Move alignment checks at the end of
parallel region.
This patch adds support for OpenMP 5.0 allocate clause for fortran. It does not
yet support the allocator-modifier as specified in OpenMP 5.1. The allocate
clause is already supported in C/C++.
This commit contains 4 following upstream commits.
69561fc781a Add support for allocate clause (OpenMP 5.0). deb9f18f677 Change kind of integer literal to fix a testcase. 6d498135016 libgomp.fortran/allocate-1.f90: Minor cleanup f62156eab7b libgomp.fortran/allocate-1.f90: Fix minor cleanup
gcc/fortran/ChangeLog:
* dump-parse-tree.c (show_omp_clauses): Handle OMP_LIST_ALLOCATE.
* gfortran.h (OMP_LIST_ALLOCATE): New enum value.
* openmp.c (enum omp_mask1): Add OMP_CLAUSE_ALLOCATE.
(gfc_match_omp_clauses): Handle OMP_CLAUSE_ALLOCATE
(OMP_PARALLEL_CLAUSES, OMP_DO_CLAUSES, OMP_SECTIONS_CLAUSES)
(OMP_TASK_CLAUSES, OMP_TASKLOOP_CLAUSES, OMP_TARGET_CLAUSES)
(OMP_TEAMS_CLAUSES, OMP_DISTRIBUTE_CLAUSES)
(OMP_SINGLE_CLAUSES): Add OMP_CLAUSE_ALLOCATE.
(OMP_TASKGROUP_CLAUSES): New.
(gfc_match_omp_taskgroup): Use OMP_TASKGROUP_CLAUSES instead of
OMP_CLAUSE_TASK_REDUCTION.
(resolve_omp_clauses): Handle OMP_LIST_ALLOCATE.
(resolve_omp_do): Avoid warning when loop iteration variable is
in allocate clause.
* trans-openmp.c (gfc_trans_omp_clauses): Handle translation of
allocate clause.
(gfc_split_omp_clauses): Update for OMP_LIST_ALLOCATE.
gcc/testsuite/ChangeLog:
* gfortran.dg/gomp/allocate-1.f90: New test.
* gfortran.dg/gomp/allocate-2.f90: New test.
* gfortran.dg/gomp/allocate-3.f90: New test.
* gfortran.dg/gomp/collapse1.f90: Update error message.
* gfortran.dg/gomp/openmp-simd-4.f90: Likewise.
* gfortran.dg/gomp/clauses-1.f90: Uncomment allocate clause.
libgomp/ChangeLog:
* testsuite/libgomp.fortran/allocate-1.c: New test.
* testsuite/libgomp.fortran/allocate-1.f90: New test.
* libgomp.texi: Remove string that says that allocate clause
support is for C/C++ only.
As Thomas reported in
https://gcc.gnu.org/pipermail/gcc-patches/2022-January/588448.html
a test added in my recent allocate clause patch fails on m32. It was due
to default kind for integer matching c_intptr_t for m32. I have now
changed it to 0_1 so that always integer with kind=1 is used.
gcc/testsuite/ChangeLog:
* gfortran.dg/gomp/allocate-2.f90: Change 0 to 0_1.
PR fortran/101762
* expr.c (gfc_check_pointer_assign): For pointer initialization
targets, check that subscripts and substring indices in
specifications are constant expressions.
gcc/testsuite/ChangeLog:
PR fortran/101762
* gfortran.dg/pr101762.f90: New test.
Tobias Burnus [Wed, 2 Mar 2022 19:02:15 +0000 (20:02 +0100)]
Fortran/OpenMP: class.cc fix for mapping of DT with allocatable components
This commit: OG11 version.
GCC 12/mainline submission (previous commit and this follow up):
https://gcc.gnu.org/pipermail/gcc-patches/2022-March/591144.html
Tobias Burnus [Tue, 1 Mar 2022 15:35:08 +0000 (16:35 +0100)]
Fortran/OpenMP: Support mapping of DT with allocatable components
This commit: OG11 version.
GCC 12/mainline submission:
https://gcc.gnu.org/pipermail/gcc-patches/2022-March/591075.html
gcc/fortran/ChangeLog:
* class.c (finalization_scalarizer): Mark syms as artificial.
(generate_callback_wrapper): New.
(gfc_find_derived_vtab): Call it, add _callback comp.
* f95-lang.c (LANG_HOOKS_OMP_DEEP_MAPPING,
LANG_HOOKS_OMP_DEEP_MAPPING_P,
LANG_HOOKS_OMP_DEEP_MAPPING_CNT): Redeinfe
* gfortran.h (gfc_import_iso_c_binding_module,
GFC_CLASS_CALLBACK_DEFAULT_FLAG, GFC_CLASS_CALLBACK_VTABLE_FLAG,
GFC_CLASS_CB_ALLOCATABLE, GFC_CLASS_CB_POINTER,
GFC_CLASS_CB_PROC_POINTER, GFC_CLASS_CB_VTABLE,
GFC_CLASS_CB_VPTR): New.
* match.c (select_type_set_tmp): Propagate allocatable property.
* module.c (MOD_VERSION): Bump due to vtab change.
(import_iso_c_binding_module): New import_all arg.
(gfc_import_iso_c_binding_module): New.
(gfc_use_module): Update call.
* openmp.c (resolve_omp_clauses): Accept DT with alloc comps.
* resolve.c (gfc_resolve_formal_arglist, gfc_resolve_intrinsic,
resolve_fl_procedure, resolve_types): Permit some violations
for internal code.
* trans-array.c (gfc_conv_descriptor_stride_get,
gfc_tree_array_size, gfc_full_array_size): Update
for GFC_TYPE_ARRAY_AKIND change.
(gfc_conv_expr_descriptor): Likewise; permit calling with tree code.
* trans-expr.c (VTABLE_CALLBACK_FIELD): Add.
(VTAB_GET_FIELD_GEN): Use it.
(VTABLE_DEALLOCATE_FIELD): Undef at the end.
(gfc_conv_expr_reference): Fixes; avoid unneccessary temp var.
* trans-intrinsic.c (gfc_conv_intrinsic_sizeof,
gfc_conv_associated): Fix class and comp-ref handling.
(conv_isocbinding_function): Remove buggy code.
* trans-openmp.c (gfc_has_alloc_comps): Add ptr_ok arg.
(gfc_omp_private_outer_ref, gfc_walk_alloc_comps,
gfc_omp_clause_default_ctor, gfc_omp_clause_copy_ctor,
gfc_omp_clause_assign_op, gfc_omp_clause_dtor,
(gfc_omp_finish_clause): Update call.
(GFC_MAP_TOKEN_DATA, GFC_MAP_TOKEN_SIZES, GFC_MAP_TOKEN_KINDS,
GFC_MAP_TOKEN_DATA_OFFSET, GFC_MAP_TOKEN_OFFSET,
GFC_MAP_TOKEN_FLAGS, GFC_MAP_TOKEN_DETACH): Define.
(gfc_omp_get_token_data, gfc_omp_get_token_sizes,
gfc_omp_get_token_kinds, gfc_omp_get_token_offset_data,
gfc_omp_get_token_offset, gfc_omp_get_token_flags,
gfc_omp_get_token_detach, gfc_omp_get_map_token_type,
gfc_omp_get_cb_type, gfc_omp_gen_deep_map_fn,
gfc_omp_deep_mapping_map, gfc_omp_deep_mapping_item,
gfc_omp_deep_mapping_comps, gfc_omp_gen_simple_loop,
gfc_omp_get_array_size, gfc_omp_elmental_loop,
gfc_omp_deep_map_kind_p, gfc_omp_deep_mapping_int_p,
gfc_omp_deep_mapping_p, gfc_omp_deep_mapping_do),
gfc_omp_deep_mapping_cnt, gfc_omp_deep_mapping): New.
(gfc_trans_omp_array_section): Save clause decl to survive gimplifying.
(gfc_trans_omp_clauses): Likewise; fixes.
* trans-types.c (gfc_build_array_type, gfc_get_derived_type,
gfc_get_array_descr_info): Update array kind to distinguish
different assumed-rank arrays.
* trans.h (gfc_class_vtab_callback_get, gfc_omp_deep_mapping_p,
gfc_omp_deep_mapping_cnt, gfc_omp_deep_mapping): New prototypes.
(enum gfc_array_kind): Additional GFC_ARRAY_ASSUMED_RANK_* entries.
gcc/ChangeLog:
* langhooks-def.h (lhd_omp_deep_mapping_p,
lhd_omp_deep_mapping_cnt, lhd_omp_deep_mapping): New.
(LANG_HOOKS_OMP_DEEP_MAPPING_P, LANG_HOOKS_OMP_DEEP_MAPPING_CNT,
LANG_HOOKS_OMP_DEEP_MAPPING): Define.
(LANG_HOOKS_DECLS): Use it.
* langhooks.c (lhd_omp_deep_mapping_p, lhd_omp_deep_mapping_cnt,
lhd_omp_deep_mapping): New stubs.
* langhooks.h (struct lang_hooks_for_decls): Add new hooks
* omp-expand.c (expand_omp_target): Handle dynamic-size
addr/sizes/kinds arrays.
* omp-low.c (build_sender_ref, fixup_child_record_type,
scan_sharing_clauses, lower_omp_target): Update to handle
new hooks and dynamic-size addr/sizes/kinds arrays.
libgomp/ChangeLog:
* testsuite/libgomp.fortran/allocatable-comp.f90: New test.
* testsuite/libgomp.fortran/map-alloc-comp-3.f90: New test.
* testsuite/libgomp.fortran/map-alloc-comp-4.f90: New test.
* testsuite/libgomp.fortran/map-alloc-comp-5.f90: New test.
* testsuite/libgomp.fortran/map-alloc-comp-6.f90: New test.
* testsuite/libgomp.fortran/map-alloc-comp-7.f90: New test.
Tobias Burnus [Tue, 15 Feb 2022 20:42:33 +0000 (21:42 +0100)]
Fortran/OpenMP: Fix depend-clause handling for c_ptr
gcc/fortran/ChangeLog:
* trans-openmp.cc (gfc_trans_omp_depobj): Fix to alloc/ptr dummy
and for c_ptr.
gcc/testsuite/ChangeLog:
* gfortran.dg/gomp/depend-4.f90: Add VALUE test, update scan test.
* gfortran.dg/gomp/depend-5.f90: Fix scan tree for -m32.
* gfortran.dg/gomp/depend-6.f90: New test.
That commit fixed an issue with omp_get_device_num() on gcn/nvptx that
resulted in having always the value 0.
This commit modifies the tests to iterate over all devices such that on a
multi-nonhost-device system it had detected that always-zero issue.
libgomp/ChangeLog:
* testsuite/libgomp.c-c++-common/target-45.c: Iterate over all devices.
* testsuite/libgomp.fortran/target10.f90: Likewise.
Tobias Burnus [Sat, 4 Dec 2021 18:39:43 +0000 (19:39 +0100)]
Fortran/OpenMP: Support most of 5.1 atomic extensions
Implements moste of OpenMP 5.1 atomic extensions,
except that 'compare' is parsed but rejected during
resolution. (As the trans-openmp.c handling is missing.)
gcc/fortran/ChangeLog:
* dump-parse-tree.c (show_omp_clauses): Handle
weak/compare/fail clause.
* gfortran.h (gfc_omp_clauses): Add weak, compare, fail.
* openmp.c (enum omp_mask1, gfc_match_omp_clauses,
OMP_ATOMIC_CLAUSES): Update for new clauses.
(gfc_match_omp_atomic): Update for 5.1 atomic changes.
(is_conversion): Support widening in one go.
(is_scalar_intrinsic_expr): New.
(resolve_omp_atomic): Update for 5.1 atomic changes.
* parse.c (parse_omp_oacc_atomic): Update for compare.
* resolve.c (gfc_resolve_blocks): Update asserts.
* trans-openmp.c (gfc_trans_omp_atomic): Handle new clauses.
gcc/testsuite/ChangeLog:
* gfortran.dg/gomp/atomic-2.f90: Move now supported code to ...
* gfortran.dg/gomp/atomic.f90: here.
* gfortran.dg/gomp/atomic-10.f90: New test.
* gfortran.dg/gomp/atomic-12.f90: New test.
* gfortran.dg/gomp/atomic-15.f90: New test.
* gfortran.dg/gomp/atomic-16.f90: New test.
* gfortran.dg/gomp/atomic-17.f90: New test.
* gfortran.dg/gomp/atomic-18.f90: New test.
* gfortran.dg/gomp/atomic-19.f90: New test.
* gfortran.dg/gomp/atomic-20.f90: New test.
* gfortran.dg/gomp/atomic-22.f90: New test.
* gfortran.dg/gomp/atomic-24.f90: New test.
* gfortran.dg/gomp/atomic-25.f90: New test.
* gfortran.dg/gomp/atomic-26.f90: New test.
Jakub Jelinek [Mon, 15 Nov 2021 12:20:53 +0000 (13:20 +0100)]
openmp: Add support for thread_limit clause on target
OpenMP 5.1 says that thread_limit clause can also appear on target,
and similarly to teams should affect the thread-limit-var ICV.
On combined target teams, the clause goes to both.
We actually passed thread_limit internally on target already before,
but only used it for gcn/ptx offloading to hint how many threads should be
created and for ptx didn't set thread_limit_var in that case.
Similarly for host fallback.
Also, I found that we weren't copying the args array that contains encoded
thread_limit and num_teams clause for target (etc.) for async target.
2021-11-15 Jakub Jelinek <jakub@redhat.com>
gcc/
* gimplify.c (optimize_target_teams): Only add OMP_CLAUSE_THREAD_LIMIT
to OMP_TARGET_CLAUSES if it isn't there already.
gcc/c-family/
* c-omp.c (c_omp_split_clauses) <case OMP_CLAUSE_THREAD_LIMIT>:
Duplicate to both OMP_TARGET and OMP_TEAMS.
gcc/c/
* c-parser.c (OMP_TARGET_CLAUSE_MASK): Add
PRAGMA_OMP_CLAUSE_THREAD_LIMIT.
gcc/cp/
* parser.c (OMP_TARGET_CLAUSE_MASK): Add
PRAGMA_OMP_CLAUSE_THREAD_LIMIT.
libgomp/
* task.c (gomp_create_target_task): Copy args array as well.
* target.c (gomp_target_fallback): Add args argument.
Set gomp_icv (true)->thread_limit_var if thread_limit is present.
(GOMP_target): Adjust gomp_target_fallback caller.
(GOMP_target_ext): Likewise.
(gomp_target_task_fn): Likewise.
* config/nvptx/team.c (gomp_nvptx_main): Set
gomp_global_icv.thread_limit_var.
* testsuite/libgomp.c-c++-common/thread-limit-1.c: New test.
Tobias Burnus [Fri, 12 Nov 2021 16:58:21 +0000 (17:58 +0100)]
Fortran/openmp: Fix '!$omp end'
gcc/fortran/ChangeLog:
* parse.c (decode_omp_directive): Fix permitting 'nowait' for some
combined directives, add missing 'omp end ... loop'.
(gfc_ascii_statement): Fix ST_OMP_END_TEAMS_LOOP result.
* openmp.c (resolve_omp_clauses): Add missing combined loop constructs
case values to the 'if(directive-name: ...)' check.
* trans-openmp.c (gfc_split_omp_clauses): Put nowait on target if
first leaf construct accepting it.
gcc/testsuite/ChangeLog:
* gfortran.dg/gomp/unexpected-end.f90: Update dg-error.
* gfortran.dg/gomp/clauses-1.f90: New test.
* gfortran.dg/gomp/nowait-2.f90: New test.
* gfortran.dg/gomp/nowait-3.f90: New test.
Jakub Jelinek [Thu, 11 Nov 2021 08:42:47 +0000 (09:42 +0100)]
openmp: Add support for 2 argument num_teams clause
In OpenMP 5.1, num_teams clause can accept either one expression as before,
but it in that case changed meaning, rather than create <= expression
teams it is now create == expression teams. Or it accepts two expressions
separated by :, with the meaning that the first is low bound and second upper
bound on how many teams should be created. The other ways to set number of
teams are upper bounds with lower bound of 1.
The following patch does parsing of this for C/C++. For host teams, we
actually don't need to do anything further right now, we always create
(pretend to create) exactly the requested number of teams, so we can just
evaluate and throw away the lower bound for now.
For teams nested in target, we don't guarantee that though and further
work will be needed.
In particular, omplower now turns the teams part of:
struct S { S (); S (const S &); ~S (); int s; };
void bar (S &, S &);
int baz ();
_Pragma ("omp declare target to (baz)");
void
foo (void)
{
S a, b;
#pragma omp target private (a) map (b)
{
#pragma omp teams firstprivate (b) num_teams (baz ())
{
bar (a, b);
}
}
}
into:
retval.0 = baz ();
retval.1 = retval.0;
{
unsigned int retval.3;
struct S * D.2549;
struct S b;
retval.3 = (unsigned int) retval.1;
D.2549 = .omp_data_i->b;
S::S (&b, D.2549);
#pragma omp teams num_teams(retval.1) firstprivate(b) shared(a)
__builtin_GOMP_teams (retval.3, 0);
{
bar (&a, &b);
}
S::~S (&b);
#pragma omp return(nowait)
}
IMHO we want a new API, say GOMP_teams3 which will take 3 arguments
instead of 2 (the lower and upper bounds from num_teams and thread_limit)
and will return a bool whether it should do the teams body or not.
And, we should add right before outermost {} above
while (__builtin_GOMP_teams3 ((unsigned) retval.1, (unsigned) retval.1, 0))
and remove the __builtin_GOMP_teams call. The current function performs
exit equivalent (at least on NVPTX) which seems bad because that means
the destructors of e.g. private variables on target aren't invoked, and
at the current placement neither destructors of the already constructed
privatized variables in teams.
I'll do this next on the compiler side, but I'm afraid I'll need help
with the nvptx and amdgcn implementations. E.g. for nvptx, we won't be
able to use %ctaid.x . I think ideal would be to use a .shared
integer variable for the omp_get_team_num value, but I don't have any
experience with that, are .shared variables zero initialized by default,
or do they have random value at start? PTX docs say they aren't initializable.
2021-11-11 Jakub Jelinek <jakub@redhat.com>
gcc/
* tree.h (OMP_CLAUSE_NUM_TEAMS_EXPR): Rename to ...
(OMP_CLAUSE_NUM_TEAMS_UPPER_EXPR): ... this.
(OMP_CLAUSE_NUM_TEAMS_LOWER_EXPR): Define.
* tree.c (omp_clause_num_ops): Increase num ops for
OMP_CLAUSE_NUM_TEAMS to 2.
* tree-pretty-print.c (dump_omp_clause): Print optional lower bound
for OMP_CLAUSE_NUM_TEAMS.
* gimplify.c (gimplify_scan_omp_clauses): Gimplify
OMP_CLAUSE_NUM_TEAMS_LOWER_EXPR if non-NULL.
(optimize_target_teams): Use OMP_CLAUSE_NUM_TEAMS_UPPER_EXPR instead
of OMP_CLAUSE_NUM_TEAMS_EXPR. Handle OMP_CLAUSE_NUM_TEAMS_LOWER_EXPR.
* omp-low.c (lower_omp_teams): Use OMP_CLAUSE_NUM_TEAMS_UPPER_EXPR
instead of OMP_CLAUSE_NUM_TEAMS_EXPR.
* omp-expand.c (expand_teams_call, get_target_arguments): Likewise.
gcc/c/
* c-parser.c (c_parser_omp_clause_num_teams): Parse optional
lower-bound and store it into OMP_CLAUSE_NUM_TEAMS_LOWER_EXPR.
Use OMP_CLAUSE_NUM_TEAMS_UPPER_EXPR instead of
OMP_CLAUSE_NUM_TEAMS_EXPR.
(c_parser_omp_target): For OMP_CLAUSE_NUM_TEAMS evaluate before
combined target teams even lower-bound expression.
gcc/cp/
* parser.c (cp_parser_omp_clause_num_teams): Parse optional
lower-bound and store it into OMP_CLAUSE_NUM_TEAMS_LOWER_EXPR.
Use OMP_CLAUSE_NUM_TEAMS_UPPER_EXPR instead of
OMP_CLAUSE_NUM_TEAMS_EXPR.
(cp_parser_omp_target): For OMP_CLAUSE_NUM_TEAMS evaluate before
combined target teams even lower-bound expression.
* semantics.c (finish_omp_clauses): Handle
OMP_CLAUSE_NUM_TEAMS_LOWER_EXPR of OMP_CLAUSE_NUM_TEAMS clause.
* pt.c (tsubst_omp_clauses): Likewise.
(tsubst_expr): For OMP_CLAUSE_NUM_TEAMS evaluate before
combined target teams even lower-bound expression.
gcc/fortran/
* trans-openmp.c (gfc_trans_omp_clauses): Use
OMP_CLAUSE_NUM_TEAMS_UPPER_EXPR instead of OMP_CLAUSE_NUM_TEAMS_EXPR.
gcc/testsuite/
* c-c++-common/gomp/clauses-1.c (bar): Supply lower-bound expression
to half of the num_teams clauses.
* c-c++-common/gomp/num-teams-1.c: New test.
* c-c++-common/gomp/num-teams-2.c: New test.
* g++.dg/gomp/attrs-1.C (bar): Supply lower-bound expression
to half of the num_teams clauses.
* g++.dg/gomp/attrs-2.C (bar): Likewise.
* g++.dg/gomp/num-teams-1.C: New test.
* g++.dg/gomp/num-teams-2.C: New test.
libgomp/
* testsuite/libgomp.c-c++-common/teams-1.c: New test.
Tobias Burnus [Sat, 30 Oct 2021 21:45:32 +0000 (23:45 +0200)]
OpenMP: Add strictly nested API call check [PR102972]
The teams construct only permits omp_get_num_teams and omp_get_team_num
as API call in strictly nested regions - check for it.
Additionally, for Fortran, using DECL_NAME does not show the mangled
name, hence, DECL_ASSEMBLER_NAME had to be used to.
Finally, 'target device(ancestor:1)' wrongly rejected non-API calls
as well.
PR middle-end/102972
gcc/ChangeLog:
* omp-low.c (omp_runtime_api_call): Use DECL_ASSEMBLER_NAME to get
internal Fortran name; new permit_num_teams arg to permit
omp_get_num_teams and omp_get_team_num.
(scan_omp_1_stmt): Update call to it, add missing call for
reverse offload, and check for strictly nested API calls in teams.
gcc/testsuite/ChangeLog:
* c-c++-common/gomp/target-device-ancestor-3.c: Add non-API
routine test.
* gfortran.dg/gomp/order-6.f90: Add missing bind(C).
* c-c++-common/gomp/teams-3.c: New test.
* gfortran.dg/gomp/teams-3.f90: New test.
* gfortran.dg/gomp/teams-4.f90: New test.
Chung-Lin Tang [Thu, 24 Feb 2022 09:07:48 +0000 (01:07 -0800)]
openmp: Handle C/C++ array reference base-pointers in array sections
In cases where a program constructs its own deep-copying for arrays-of-pointers,
e.g:
#pragma omp target enter data map(to:level->vectors[:N])
for (i = 0; i < N; i++)
#pragma omp target enter data map(to:level->vectors[i][:N])
We need to treat the part of the array reference before the array section
as a base-pointer (here 'level->vectors[i]'), providing pointer-attachment
behavior.
This patch adds this inside handle_omp_array_sections(), tracing the whole
sequence of array dimensions, creating a whole base-pointer reference
iteratively using build_array_ref(). The conditions are that each of the
"absorbed" dimensions must be length==1, and the final reference must be
of pointer-type (so that pointer attachment makes sense).
openmp: Improve handling of nested OpenMP metadirectives in C and C++
This patch fixes a misparsing issue when encountering code like:
#pragma omp metadirective when {<selector_set>={...}: A)
#pragma omp metadirective when (<selector_set>={...}: B)
When called for the first metadirective, analyze_metadirective_body would
stop just before the colon in the second metadirective because it naively
assumes that the '}' marks the end of a code block.
The assertion for clauses to end parsing at the same point is now disabled
if a parse error has occurred during the parsing of the clause, since some
tokens may not be consumed if a parse error cuts parsing short.
gcc/c/
* c-parser.c (c_parser_omp_construct): Move handling of
PRAGMA_OMP_METADIRECTIVE from here...
(c_parser_pragma): ...to here.
(analyze_metadirective_body): Check that the bracket nesting level
is also zero before stopping the adding of tokens on encountering a
close brace.
(c_parser_omp_metadirective): Modify function signature and update.
Do not assert on remaining tokens if there has been a parse error.
gcc/cp/
* parser.c (cp_parser_omp_construct): Move handling of
PRAGMA_OMP_METADIRECTIVE from here...
(cp_parser_pragma): ...to here.
(analyze_metadirective_body): Check that the bracket
nesting level is also zero before stopping the adding of tokens on
encountering a close brace.
(cp_parser_omp_metadirective): Modify function signature and update.
Do not assert on remaining tokens if there has been a parse error.
gcc/testsuite/
* c-c++-common/gomp/metadirective-1.c (f): Add test for
improperly nested metadirectives.
openmp: More Fortran front-end fixes for metadirectives
This adds a check for declarative OpenMP directives in metadirective
variants (already present in the C/C++ front-ends), and fixes an
ICE when an empty metadirective (i.e. just '!$omp metadirective')
is presented.
gcc/fortran/
* gfortran.h (is_omp_declarative_stmt): New.
* openmp.c (match_omp_metadirective): Reject declarative OpenMP
directives with 'sorry'.
* parse.c (parse_omp_metadirective_body): Check that state stack head
is non-null before dereferencing.
(is_omp_declarative_stmt): New.
gcc/testsuite/
* gfortran.dg/gomp/metadirective-2.f90 (main): Test empty
metadirective.
openmp: Eliminate non-matching metadirective variants early in Fortran front-end
This patch checks during parsing if a metadirective selector is both
resolvable and non-matching - if so, it is removed from further
consideration. This is both more efficient, and avoids spurious
syntax errors caused by considering combinations of selectors that
lead to invalid combinations of OpenMP directives, when that
combination would never arise in the first place.
This exposes another bug - when metadirectives that are not of the
begin-end variety are nested, we might have to drill up through
multiple layers of the state stack to reach the state for the
next statement. This is now fixed.
gcc/
* omp-general.c (DELAY_METADIRECTIVES_AFTER_LTO): Check that cfun is
non-null before derefencing.
gcc/fortran/
* decl.c (gfc_match_end): Search for first previous state that is not
COMP_OMP_METADIRECTIVE.
* gfortran.h (gfc_skip_omp_metadirective_clause): Add prototype.
* openmp.c (match_omp_metadirective): Skip clause if
result of gfc_skip_omp_metadirective_clause is true.
* trans-openmp.c (gfc_trans_omp_set_selector): Add argument and
disable expression conversion if false.
(gfc_skip_omp_metadirective_clause): New.
With the OpenMP 5.0/5.1 specifications, we would expect A to be selected in
the metadirective when the target call is made, but B when f is called
directly outside of a target context. However, since GCC does not have
separate copies of f for local and target calls, and the construct selector
is static, it must be resolved one way or the other at compile-time (currently
in the favour of selecting A), which may be unexpected behaviour.
This patch attempts to detect the above situation, and will emit a warning
if found.
gcc/
* gimplify.c (gimplify_omp_metadirective): Mark offloadable functions
containing metadirectives with 'construct={target}' in the selector.
* omp-general.c (omp_has_target_constructor_p): New.
* omp-general.h (omp_has_target_constructor_p): New prototype.
* omp-low.c (lower_omp_1): Emit warning if marked functions called
outside of a target context.
In gimplify_scan_omp_clauses, the case for handling indirect accesses (which
creates firstprivate ptr and zero-length array section map for such decls) was
erroneously went into for non-pointer cases (here being the base struct decl),
so added the
appropriate checks there.
Added new testcase is a compile only test for the ICE. The original omptests
t-partial-struct test actually should not execute correctly, because for
map(t.s->a[:N]), map(t.s[:1]) is not implicitly mapped, thus the entire
offloaded access does not work as is (fixing that omptests test is out of
scope here).
openmp: Add support for streaming metadirectives and resolving them after LTO
This patch adds support for streaming metadirective Gimple statements during
LTO, and adds a metadirective expansion pass that runs after LTO. This is
required for metadirectives with selectors that can only be resolved from
within the accel compiler.
openmp: Add support for resolving metadirectives during parsing and Gimplification
This adds support for resolving metadirectives according to the OpenMP 5.1
specification. The variants are sorted by score, then gathered into a list
of dynamic replacement candidates. The metadirective is then expanded into
a sequence of 'if..else' statements to test the dynamic selector and execute
the variant if the selector is satisfied.
If any of the selectors in the list are unresolvable, GCC will give up on
resolving the metadirective and try again later.
This adds a new Gimple statement type GIMPLE_OMP_METADIRECTIVE, which
represents the metadirective in Gimple. In high Gimple, the statement
contains the body of the directive variants, whereas in low Gimple, it
only contains labels to the bodies.
This patch adds support for converting metadirectives from tree to Gimple
form, and handling of the Gimple form (Gimple lowering, OpenMP lowering
and expansion, inlining, SSA handling etc).
Metadirectives should be resolved before they reach the back-end, otherwise
the compiler will crash as GCC does not know how to convert metadirective
Gimple statements to RTX.
This patch implements parsing for the OpenMP metadirective introduced in
OpenMP 5.0. Metadirectives are parsed into an OMP_METADIRECTIVE node,
with the variant clauses forming a chain accessible via
OMP_METADIRECTIVE_CLAUSES. Each clause contains the context selector
and tree for the variant.
User conditions in the selector are now permitted to be non-constant when
used in metadirectives as specified in OpenMP 5.1.
gcc/
* omp-general.c (omp_context_selector_matches): Add extra argument.
(omp_resolve_metadirective): New stub function.
* omp-general.h (struct omp_metadirective_variant): New.
(omp_context_selector_matches): Add extra argument.
(omp_resolve_metadirective): New prototype.
* tree.def (OMP_METADIRECTIVE): New.
* tree.h (OMP_METADIRECTIVE_CLAUSES): New macro.
gcc/c/
* c-parser.c (c_parser_skip_to_end_of_block_or_statement): Handle
parentheses in statement.
(c_parser_omp_metadirective): New prototype.
(c_parser_omp_context_selector): Add extra argument. Allow
non-constant expressions.
(c_parser_omp_context_selector_specification): Add extra argument and
propagate it to c_parser_omp_context_selector.
(analyze_metadirective_body): New.
(c_parser_omp_metadirective): New.
(c_parser_omp_construct): Handle PRAGMA_OMP_METADIRECTIVE.
gcc/c-family
* c-common.h (enum c_omp_directive_kind): Add C_OMP_DIR_META.
(c_omp_expand_metadirective): New prototype.
* c-gimplify.c (genericize_omp_metadirective_stmt): New.
(c_genericize_control_stmt): Handle OMP_METADIRECTIVE tree nodes.
* c-omp.c (omp_directives): Classify metadirectives as C_OMP_DIR_META.
(c_omp_expand_metadirective): New stub function.
* c-pragma.c (omp_pragmas): Add entry for metadirective.
* c-pragma.h (enum pragma_kind): Add PRAGMA_OMP_METADIRECTIVE.
Sandra Loosemore [Wed, 19 Jan 2022 20:50:49 +0000 (12:50 -0800)]
Fortran: Fix scope for OMP AFFINITY clause iterator variables [PR103695]
gfc_finish_var_decl was confused by the undocumented overloading of
the proc_name field in struct gfc_namespace to contain iterator
variables for the OpenMP AFFINITY clause, causing it to insert the
decls in the wrong scope. This patch adds a new distinct field to
hold these variables.
libgomp, OpenMP: Fix issue for omp_get_device_num on gcn targets.
Currently omp_get_device_num does not work on gcn targets with more than one
offload device. The reason is that GOMP_DEVICE_NUM_VAR is static in
icv-device.c and thus "__gomp_device_num" is not visible in the offload image.
This patch removes "static" such that "__gomp_device_num" is now part of the
offload image and can now be found in GOMP_OFFLOAD_load_image in the plugin.
This is not an issue for nvptx. There, "__gomp_device_num" is in the offload
image even with "static".
libgomp/ChangeLog:
* config/gcn/icv-device.c: Make GOMP_DEVICE_NUM_VAR public (remove
"static") to make the device num available in the offload image.
Chung-Lin Tang [Tue, 4 Jan 2022 09:26:23 +0000 (17:26 +0800)]
libgomp: Fix GOMP_DEVICE_NUM_VAR stringification during offload image load
In the patch that implemented omp_get_device_num(), there was an error where
the stringification of GOMP_DEVICE_NUM_VAR, which is the macro expanding to
the actual symbol used, was erroneously using the STRINGX() macro in the
libgomp offload image symbol search, and expansion of the variable name
string through the additional layer of preprocessor symbol was not properly
achieved.
This patch fixes this by changing to properly use XSTRING(), also from
include/symcat.h.
libgomp/ChangeLog:
* plugin/plugin-gcn.c (GOMP_OFFLOAD_load_image): Change uses of STRINGX
into XSTRING when looking for GOMP_DEVICE_NUM_VAR in offload image.
* plugin/plugin-nvptx.c (GOMP_OFFLOAD_load_image): Likewise.
Chung-Lin Tang [Tue, 4 Jan 2022 07:37:15 +0000 (15:37 +0800)]
openmp: Fix ICE in gimplify_omp_affinity [PR103643]
After the PR90030 patch, which removes the universal casting of all Fortran
array pointers to 'c_char*', a Fortran descriptor based array passed into an
affinity() clause now looks like:
The 'integer(kind=4)[0:]' incomplete type appears to be causing ICE during
gimplify_expr() due to 'is_gimple_val, fb_rvalue'. The ICE appears to be fixed
just by adjusting to 'is_gimple_lvalue, fb_lvalue'. Considering the use of the
affinity() clause, which should be specifying the location of a particular
object in memory, this probably makes sense.
gcc/ChangeLog:
PR middle-end/103643
* gimplify.c (gimplify_omp_affinity): Adjust gimplify_expr of entire
OMP_CLAUSE_DECL to use 'is_gimple_lvalue, fb_lvalue'
Andrew Stubbs [Tue, 16 Nov 2021 10:32:35 +0000 (10:32 +0000)]
amdgcn: Change offload variable table discovery
Up to now the libgomp GCN plugin has been finding the offload variables
by using a symbol lookup, but the AMD runtime requires that the symbols are
global for that to work. This was ensured by mkoffload as a post-procssing
step, but the LLVM 13 assembler no longer accepts this in the case where the
variable was previously declared differently.
This patch switches to locating the symbols directly from the
offload_var_table, which means that only one symbol needs to be forced
global.
This changes breaks the libgomp image compatibility so GOMP_VERSION_GCN has
also been bumped.
gcc/ChangeLog:
* config/gcn/mkoffload.c (process_asm): Process the variable table
completely differently.
(process_obj): Encode the varaible data differently.
include/ChangeLog:
* gomp-constants.h (GOMP_VERSION_GCN): Bump.
libgomp/ChangeLog:
* plugin/plugin-gcn.c (struct gcn_image_desc): Remove global_variables.
(GOMP_OFFLOAD_load_image): Locate the offload variables via the
table, not individual symbols.
Andrew Stubbs [Fri, 3 Dec 2021 17:46:41 +0000 (17:46 +0000)]
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 the minimum version
requirement is now bumped to 4.1 (still old at this point).
libgomp/ChangeLog:
* allocator.c (MEMSPACE_ALLOC): New macro.
(MEMSPACE_CALLOC): New macro.
(MEMSPACE_REALLOC): New macro.
(MEMSPACE_FREE): New macro.
(dynamic_smem_size): New constants.
(omp_alloc): Use MEMSPACE_ALLOC.
Implement fall-backs for predefined allocators.
(omp_free): Use MEMSPACE_FREE.
(omp_calloc): Use MEMSPACE_CALLOC.
Implement fall-backs for predefined allocators.
(omp_realloc): Use MEMSPACE_REALLOC.
Implement fall-backs for predefined allocators.
* config/nvptx/team.c (__nvptx_lowlat_heap_root): New variable.
(__nvptx_lowlat_pool): New asm varaible.
(gomp_nvptx_main): Initialize the low-latency heap.
* 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.
* config/nvptx/allocator.c: New file.
* testsuite/libgomp.c/allocators-1.c: New test.
* testsuite/libgomp.c/allocators-2.c: New test.
* testsuite/libgomp.c/allocators-3.c: New test.
* testsuite/libgomp.c/allocators-4.c: New test.
* testsuite/libgomp.c/allocators-5.c: New test.
* testsuite/libgomp.c/allocators-6.c: New test.
Andrew Stubbs [Thu, 16 Dec 2021 15:30:05 +0000 (15:30 +0000)]
OpenMP: allow requires dynamic_allocators
There's no need to reject the dynamic_allocators requires directive because
we actually do support the feature, and it doesn't have to actually "do"
anything.
Fix issue with the Fortran front-end when mapping arrays: when creating the
data MEM_REF for the map clause, there was a convention of casting the
referencing pointer to 'c_char *' by
fold_convert (build_pointer_type (char_type_node), ptr).
This causes the alignment passed to the libgomp runtime for array data
hardwared to '1', and causes alignment errors on the offload target.
This patch fixes this by removing the char_type_node pointer converts, and
adding gcc_asserts to ensure POINTER_TYPE_P (TREE_TYPE (ptr)).
PR fortran/90030
gcc/fortran/ChangeLog:
* trans-openmp.c (gfc_omp_finish_clause): Remove fold_convert to pointer
to char_type_node, add gcc_assert of POINTER_TYPE_P.
(gfc_trans_omp_array_section): Likewise.
(gfc_trans_omp_clauses): Likewise.
Chung-Lin Tang [Fri, 3 Dec 2021 09:27:17 +0000 (17:27 +0800)]
fortran: Fix setting of array lower bound for named arrays
This patch fixes a case of setting array low-bounds, found for particular uses
of SOURCE=/MOLD=. This adjusts the relevant part in gfc_trans_allocate() to
set e3_has_nodescriptor only for non-named arrays.
Frederik Harwath [Tue, 16 Nov 2021 15:22:29 +0000 (16:22 +0100)]
graphite: Accept loops without data references
It seems that the check that rejects loops without data references is
only included to avoid handling non-profitable loops. Including those
loops in Graphite's analysis enables more consistent diagnostic
messages in OpenACC "kernels" code and does not introduce any
testsuite regressions. If executing Graphite on loops without
data references leads to noticeable compile time slow-downs for
non-OpenACC users of Graphite, the check can be re-introduced but
restricted to non-OpenACC functions.
gcc/ChangeLog:
* graphite-scop-detection.c (scop_detection::harmful_loop_in_region):
Remove check for loops without data references.
Frederik Harwath [Tue, 16 Nov 2021 15:21:57 +0000 (16:21 +0100)]
graphite: Adjust scop loop-nest choice
The find_common_loop function is used in Graphite to obtain a common
super-loop of all loops inside a SCoP. The function is applied to the
loop of the destination block of the edge that leads into the SESE
region and the loop of the source block of the edge that exits the
region. The exit block is usually introduced by the canonicalization
of the loop structure that Graphite does to support its code
generation. If it is empty, it may happen that it belongs to the outer
fake loop. This way, build_alias_set may end up analysing
data-references with respect to this loop although there may exist a
proper super-loop of the SCoP loops. This does not seem to be correct
in general and it leads to problems with runtime alias check creation
which fails if executed on a loop without niter information.
gcc/ChangeLog:
* graphite-scop-detection.c (scop_context_loop): New function.
(build_alias_set): Use scop_context_loop instead of find_common_loop.
* graphite-isl-ast-to-gimple.c (graphite_regenerate_ast_isl): Likewise.
* graphite.h (scop_context_loop): New declaration.
Frederik Harwath [Tue, 16 Nov 2021 15:21:42 +0000 (16:21 +0100)]
graphite: Tune parameters for OpenACC use
The default values of some parameters that restrict Graphite's
resource usage are too low for many OpenACC codes. Furthermore,
exceeding the limits does not alwas lead to user-visible diagnostic
messages.
This commit increases the parameter values on OpenACC functions. The
values were chosen to allow for the analysis of all "kernels" regions
in the SPEC ACCEL v1.3 benchmark suite. Warnings about exceeded
Graphite-related limits are added to the -fopt-info-missed
output. Those warnings are phrased in a uniform way that intentionally
refers to the "data-dependence analysis" of "OpenACC loops" instead of
"a failure in Graphite" to make them easier to understand for users.
gcc/ChangeLog:
* graphite-optimize-isl.c (optimize_isl): Adjust
param_max_isl_operations value for OpenACC functions and add
special warnings if value gets exceeded.
* graphite-scop-detection.c (build_scops): Likewise for
param_graphite_max_arrays_per_scop.
gcc/testsuite/ChangeLog:
* gcc.dg/goacc/graphite-parameter-1.c: New test.
* gcc.dg/goacc/graphite-parameter-2.c: New test.
Frederik Harwath [Tue, 16 Nov 2021 15:20:56 +0000 (16:20 +0100)]
openacc: Disable pass_pre on outlined functions analyzed by Graphite
The additional dependences introduced by partial redundancy
elimination proper and by the code hoisting step of the pass very
often cause Graphite to fail on OpenACC functions. On the other hand,
the pass can also enable the analysis of OpenACC loops (cf. e.g. the
loop-auto-transfer-4.f90 testcase), for instance, because full
redundancy elimination removes definitions that would otherwise
prevent the creation of runtime alias checks outside of the SCoP.
This commit disables the actual partial redundancy elimination step as
well as the code hoisting step of pass_pre on OpenACC functions that
might be handled by Graphite.
gcc/ChangeLog:
* tree-ssa-pre.c (insert): Skip any insertions in OpenACC
functions that might be processed by Graphite.
Frederik Harwath [Tue, 16 Nov 2021 15:20:41 +0000 (16:20 +0100)]
openacc: Handle internal function calls in pass_lim
The loop invariant motion pass correctly refuses to move statements
out of a loop if any other statement in the loop is unanalyzable. The
pass does not know how to handle the OpenACC internal function calls
which was not necessary until recently when the OpenACC device
lowering pass was moved to a later position in the pass pipeline.
This commit changes pass_lim to ignore the OpenACC internal function
calls which do not contain any memory references. The hoisting enabled
by this change can be useful for the data-dependence analysis in
Graphite; for instance, in the outlined functions for OpenACC regions,
all invariant accesses to the ".omp_data_i" struct should be hoisted
out of the OpenACC loop. This is particularly important for variables
that were scalars in the original loop and which have been turned into
accesses to the struct by the outlining process. Not hoisting those
can prevent scalar evolution analysis which is crucial for Graphite.
Since any hoisting that introduces intermediate names - and hence,
"fake" dependences - inside the analyzed nest can be harmful to
data-dependence analysis, a flag to restrict the hoisting in OpenACC
functions is added to the pass. The pass instance that executes before
Graphite now runs with this flag set to true and the pass instance
after Graphite runs unrestricted.
A more precise way of selecting the statements for which hoisting
should be enabled is left for a future improvement.
gcc/ChangeLog:
* passes.def: Set restrict_oacc_hoisting to true for the early
pass_lim instance.
* tree-ssa-loop-im.c (movement_possibility): Add
restrict_oacc_hoisting flag to function; restrict movement if set.
(compute_invariantness): Add restrict_oacc_hoisting flag and pass it on.
(gather_mem_refs_stmt): Skip IFN_GOACC_LOOP and IFN_UNIQUE
calls.
(loop_invariant_motion_in_fun): Add restrict_oacc_hoisting flag and
pass it on.
(pass_lim::execute): Pass on new flags.
* tree-ssa-loop-manip.h (loop_invariant_motion_in_fun): Adjust declaration.
* gimple-loop-interchange.cc (pass_linterchange::execute): Adjust call to
loop_invariant_motion_in_fun.
Frederik Harwath [Tue, 16 Nov 2021 15:20:15 +0000 (16:20 +0100)]
openacc: Warn about "independent" "kernels" loops with data-dependences
This commit concerns loops in OpenACC "kernels" region that have been marked
up with an explicit "independent" clause by the user, but for which Graphite
found data dependences. A discussion on the private internal OpenACC mailing
list suggested that warning the user about the dependences woud be a more
acceptable solution than reverting the user's decision. This behavior is
implemented by the present commit.
gcc/ChangeLog:
* common.opt: Add flag Wopenacc-false-independent.
* omp-offload.c (oacc_loop_warn_if_false_independent): New function.
(oacc_loop_fixed_partitions): Call from here.
Andrew Stubbs [Tue, 16 Nov 2021 15:19:53 +0000 (16:19 +0100)]
openacc: Add runtime alias checking for OpenACC kernels
This commit adds the code generation for the runtime alias checks for
OpenACC loops that have been analyzed by Graphite. The runtime alias
check condition gets generated in Graphite. It is evaluated by the
code generated for the IFN_GOACC_LOOP internal function calls. If
aliasing is detected at runtime, the execution dimensions get adjusted
to execute the affected loops sequentially.
gcc/ChangeLog:
* graphite-isl-ast-to-gimple.c: Include internal-fn.h.
(graphite_oacc_analyze_scop): Implement runtime alias checks.
* omp-expand.c (expand_oacc_for): Add an additional "noalias" parameter
to GOACC_LOOP internal calls, and initialise it to integer_one_node.
* omp-offload.c (oacc_xform_loop): Integrate the runtime alias check
into the GOACC_LOOP expansion.
libgomp/ChangeLog:
* testsuite/libgomp.oacc-c-c++-common/runtime-alias-check-1.c: New test.
* testsuite/libgomp.oacc-c-c++-common/runtime-alias-check-2.c: New test.
Andrew Stubbs [Tue, 16 Nov 2021 15:19:23 +0000 (16:19 +0100)]
openacc: Add data optimization pass
Address PR90591 "Avoid unnecessary data transfer out of OMP
construct", for simple (but common) cases.
This commit adds a pass that optimizes data mapping clauses.
Currently, it can optimize copy/map(tofrom) clauses involving scalars
to copyin/map(to) and further to "private". The pass is restricted
"kernels" regions but could be extended to other types of regions.
gcc/ChangeLog:
* Makefile.in: Add pass.
* doc/gimple.texi: TODO.
* gimple-walk.c (walk_gimple_seq_mod): Adjust for backward walking.
* gimple-walk.h (struct walk_stmt_info): Add field.
* passes.def: Add new pass.
* tree-pass.h (make_pass_omp_data_optimize): New declaration.
* omp-data-optimize.cc: New file.
Frederik Harwath [Tue, 16 Nov 2021 15:18:02 +0000 (16:18 +0100)]
Add function for printing a single OMP_CLAUSE
Commit 89f4f339130c ("For 'OMP_CLAUSE' in 'dump_generic_node', dump
the whole OMP clause chain") changed the dumping behavior for
OMP_CLAUSEs. The old behavior is required for a follow-up
commit ("openacc: Add data optimization pass") that optimizes single
OMP_CLAUSEs.
Frederik Harwath [Tue, 16 Nov 2021 15:17:48 +0000 (16:17 +0100)]
openacc: Remove unused partitioning in "kernels" regions
With the old "kernels" handling, unparallelized regions would
get executed with 1x1x1 partitioning even if the user provided
explicit num_gangs, num_workers clauses etc.
This commit restores this behavior by removing unused partitioning
after assigning the parallelism dimensions to loops.
gcc/ChangeLog:
* omp-offload.c (oacc_remove_unused_partitioning): New function
for removing partitioning that is not used by any loop.
(oacc_validate_dims): Call oacc_remove_unused_partitioning and
enable warnings about unused partitioning.
projects / thirdparty / gcc.git / log
