mirror of
https://gcc.gnu.org/git/gcc.git
synced 2026-02-21 19:35:28 -05:00
723b18ce3d
This patch implements "managed" memory for AMD GCN GPUs in OpenMP. It builds on the support added to the NVPTX libgomp for CUDA Managed Memory, a week or two ago. These features were first posted here a few years ago, as part of a larger Unified Shared Memory patch series, and then in a slightly changed version just over a year ago. Hopefully this time the controversial bits have been removed. Since we do not use HIP we cannot use hipMallocManaged, so this patch attempts to replicate the same effect by setting the appropriate attributes. This works on more devices than support proper USM, but still I cannot be sure that the settings are correct for every device out there (I have tested on gfx900, gfx906, gfx908, gfx90a, and gfx1100). The HSA header file update uses the most recent files relicensed for us by AMD, at the time of the first patch posting. Those files have certainly moved on in the upstream sources, but I did not ask to get those relicensed. include/ChangeLog: * hsa.h: Import newer version. * hsa_ext_amd.h: Likewise. * hsa_ext_image.h: Likewise. libgomp/ChangeLog: * Makefile.in: Regenerate. * libgomp-plugin.h (gomp_simple_alloc_init_context): New prototype. (gomp_simple_alloc_register_memory): New prototype. (gomp_simple_alloc): New prototype. (gomp_simple_free): New prototype. (gomp_simple_realloc): New prototype. * libgomp.h (gomp_simple_alloc_init_context): Move to libgomp-plugin.h. (gomp_simple_alloc_register_memory): Likewise. (gomp_simple_alloc): Likewise. (gomp_simple_free): Likewise. (gomp_simple_realloc): Likewise. * libgomp.texi: Update AMD managed memory description. * plugin/Makefrag.am (libgomp_plugin_gcn_la_SOURCES): Add simple-allocator.c and plugin/mutex.c. * plugin/plugin-gcn.c: Include sys/mman.h and unistd.h. (struct hsa_runtime_fn_info): Add hsa_amd_svm_attributes_set_fn. (dump_hsa_system_info): Add HSA_AMD_SYSTEM_INFO_SVM_SUPPORTED and HSA_AMD_SYSTEM_INFO_SVM_ACCESSIBLE_BY_DEFAULT to the GCN_DEBUG output. (init_hsa_runtime_functions): Add hsa_amd_svm_attributes_set. (isa_matches_agent): Add a new error message for the case where the ISA doesn't match but the name does. (managed_ctx): New variable. (managed_heap_create): New function. (GOMP_OFFLOAD_managed_alloc): Likewise. (GOMP_OFFLOAD_managed_free): Likewise. * simple-allocator.c (gomp_fatal): New macro. * testsuite/lib/libgomp.exp (check_effective_target_omp_managedmem): Add amdgcn support checker. (check_effective_target_offload_target_amdgcn_with_xnack): New. * testsuite/libgomp.c-c++-common/requires-4.c: Ignore xnack warning. * testsuite/libgomp.c-c++-common/requires-4a.c: Ignore xnack warning. * testsuite/libgomp.c-c++-common/requires-5.c: Ignore xnack warning. * testsuite/libgomp.c++/alloc-managed-1.C: Add -mxnack=on, if needed. * testsuite/libgomp.c/alloc-managed-1.c: Likewise. * testsuite/libgomp.c/alloc-managed-2.c: Likewise. * testsuite/libgomp.c/alloc-managed-3.c: Likewise. * testsuite/libgomp.c/alloc-managed-4.c: Likewise. * testsuite/libgomp.fortran/alloc-managed-1.f90: Likewise. * plugin/mutex.c: New file.
41 lines
1.4 KiB
C
41 lines
1.4 KiB
C
/* { dg-require-effective-target lto } */
|
|
/* { dg-additional-options "-flto" } */
|
|
/* { dg-additional-options "-foffload-options=nvptx-none=-misa=sm_35" { target { offload_target_nvptx } } } */
|
|
/* { dg-additional-sources requires-4-aux.c } */
|
|
|
|
/* GCC explicitly disables XNACK for gfx908 (and others) as the hardware
|
|
support is limited, which results in a diagnostic. */
|
|
/* { dg-excess-errors "Unified Shared Memory is enabled, but XNACK is disabled" { target offload_target_amdgcn } } */
|
|
|
|
/* Check no diagnostic by device-compiler's or host compiler's lto1.
|
|
Other file uses: 'requires reverse_offload', but that's inactive as
|
|
there are no declare target directives, device constructs nor device routines */
|
|
|
|
/* Depending on offload device capabilities, it may print something like the
|
|
following (only) if GOMP_DEBUG=1:
|
|
"devices present but 'omp requires unified_address, unified_shared_memory, reverse_offload' cannot be fulfilled"
|
|
and in that case does host-fallback execution.
|
|
|
|
No offload devices support USM at present, so we may verify host-fallback
|
|
execution by presence of separate memory spaces. */
|
|
|
|
#pragma omp requires unified_address,unified_shared_memory
|
|
|
|
int a[10] = { 0 };
|
|
extern void foo (void);
|
|
|
|
int
|
|
main (void)
|
|
{
|
|
#pragma omp target map(to: a)
|
|
for (int i = 0; i < 10; i++)
|
|
a[i] = i;
|
|
|
|
for (int i = 0; i < 10; i++)
|
|
if (a[i] != i)
|
|
__builtin_abort ();
|
|
|
|
foo ();
|
|
return 0;
|
|
}
|