Commit Graph
6 Commits
Author SHA1 Message Date
Ulrich Weigand c9062e8f78 Reapply [libomptarget] Build plugins-nextgen for SystemZ (#83978)
The plugin was not getting built as the build_generic_elf64 macro
assumes the LLVM triple processor name matches the CMake processor name,
which is unfortunately not the case for SystemZ.

Fix this by providing two separate arguments instead.

Actually building the plugin exposed a number of other issues causing
various test failures. Specifically, I've had to add the SystemZ target
to
- CompilerInvocation::ParseLangArgs
- linkDevice in ClangLinuxWrapper.cpp
- OMPContext::OMPContext (to set the device_kind_cpu trait)
- LIBOMPTARGET_ALL_TARGETS in libomptarget/CMakeLists.txt
- a check_plugin_target call in libomptarget/src/CMakeLists.txt

Finally, I've had to set a number of test cases to UNSUPPORTED on
s390x-ibm-linux-gnu; all these tests were already marked as UNSUPPORTED
for x86_64-pc-linux-gnu and aarch64-unknown-linux-gnu and are failing on
s390x for what seem to be the same reason.

In addition, this also requires support for BE ELF files in
plugins-nextgen: https://github.com/llvm/llvm-project/pull/85246
2024-03-15 19:06:43 +01:00
Ulrich Weigand 70677c81de Revert "[libomptarget] Build plugins-nextgen for SystemZ (#83978)"
This reverts commit 3ecd38c8e1.
2024-03-06 21:37:43 +01:00
Ulrich Weigand 3ecd38c8e1 [libomptarget] Build plugins-nextgen for SystemZ (#83978)
The plugin was not getting built as the build_generic_elf64 macro
assumes the LLVM triple processor name matches the CMake processor name,
which is unfortunately not the case for SystemZ.

Fix this by providing two separate arguments instead.

Actually building the plugin exposed a number of other issues causing
various test failures. Specifically, I've had to add the SystemZ target
to
- CompilerInvocation::ParseLangArgs
- linkDevice in ClangLinuxWrapper.cpp
- OMPContext::OMPContext (to set the device_kind_cpu trait)
- LIBOMPTARGET_ALL_TARGETS in libomptarget/CMakeLists.txt
- a check_plugin_target call in libomptarget/src/CMakeLists.txt

Finally, I've had to set a number of test cases to UNSUPPORTED on
s390x-ibm-linux-gnu; all these tests were already marked as UNSUPPORTED
for x86_64-pc-linux-gnu and aarch64-unknown-linux-gnu and are failing on
s390x for what seem to be the same reason.

In addition, this also requires support for BE ELF files in
plugins-nextgen: https://github.com/llvm/llvm-project/pull/83976
2024-03-06 20:50:01 +01:00
carlobertolliandThorsten Blass ae99966a27 [OpenMP] Enable automatic unified shared memory on MI300A. (#77512)
This patch enables applications that did not request OpenMP
unified_shared_memory to run with the same zero-copy behavior, where
mapped memory does not result in extra memory allocations and memory
copies, but CPU-allocated memory is accessed from the device. The name
for this behavior is "automatic zero-copy" and it relies on detecting:
that the runtime is running on a MI300A, that the user did not select
unified_shared_memory in their program, and that XNACK (unified memory
support) is enabled in the current GPU configuration. If all these
conditions are met, then automatic zero-copy is triggered.

This patch also introduces an environment variable OMPX_APU_MAPS that,
if set, triggers automatic zero-copy also on non APU GPUs (e.g., on
discrete GPUs).
This patch is still missing support for global variables, which will be
provided in a subsequent patch.

Co-authored-by: Thorsten Blass <thorsten.blass@amd.com>
2024-01-22 10:30:22 -06:00
carlobertolli ce4144406c Revert "[OpenMP][libomptarget] Enable automatic unified shared memory executi…" (#77371)
Reverts llvm/llvm-project#75999

lit test is failing.
2024-01-08 14:38:29 -06:00
carlobertolliandThorsten Blass 22a73e7c46 [OpenMP][libomptarget] Enable automatic unified shared memory executi… (#75999)
…on (zero-copy) on MI300A.

This patch enables applications that did not request OpenMP
unified_shared_memory to run with the same zero-copy behavior, where
mapped memory does not result in extra memory allocations and memory
copies, but CPU-allocated memory is accessed from the device. The name
for this behavior is "automatic zero-copy" and it relies on detecting:
that the runtime is running on a MI300A, that the user did not select
unified_shared_memory in their program, and that XNACK (unified memory
support) is enabled in the current GPU configuration. If all these
conditions are met, then automatic zero-copy is triggered.

This patch is still missing support for global variables, which will be
provided in a subsequent patch.

Co-authored-by: Thorsten Blass <thorsten.blass@amd.com>
2024-01-08 14:17:28 -06:00