llvm-project.git/offload/DeviceRTL/include/DeviceUtils.h, branch users/nico/python-2 Unnamed repository; edit this file 'description' to name the repository. [OpenMP] Port the OpenMP device runtime to direct C++ compilation (#123673) 2025-02-05T14:18:52+00:00 Joseph Huber huberjn@outlook.com 2025-02-05T14:18:52+00:00 bb7ab2557c485e004e619570cca7e2204b98a71b Summary: This removes the use of OpenMP offloading to build the device runtime. The main benefit here is that we no longer need to rely on offloading semantics to build a device only runtime. Things like variants are now no longer needed and can just be simple if-defs. In the future, I will remove most of the special handling here and fold it into calls to the `<gpuintrin.h>` functions instead. Additionally I will rework the compilation to make this a separate runtime. The current plan is to have this, but make including OpenMP and offloading either automatically add it, or print a warning if it's missing. This will allow us to use a normal CMake workflow and delete all the weird 'lets pull the clang binary out of the build' business. ``` -DRUNTIMES_amdgcn-amd-amdhsa_LLVM_ENABLE_RUNTIMES=offload -DLLVM_RUNTIME_TARGETS=amdgcn-amd-amdhsa ``` After that, linking the OpenMP device runtime will be `-Xoffload-linker -lomp`. I.e. no more fat binary business. Only look at the most recent commit since this includes the two dependencies (fix to AMDGPUEmitPrintfBinding and the PointerToMember bug). 
Summary:
This removes the use of OpenMP offloading to build the device runtime.
The main benefit here is that we no longer need to rely on offloading
semantics to build a device only runtime. Things like variants are now
no longer needed and can just be simple if-defs. In the future, I will
remove most of the special handling here and fold it into calls to the
`<gpuintrin.h>` functions instead. Additionally I will rework the
compilation to make this a separate runtime.

The current plan is to have this, but make including OpenMP and
offloading either automatically add it, or print a warning if it's
missing. This will allow us to use a normal CMake workflow and delete
all the weird 'lets pull the clang binary out of the build' business.
```
-DRUNTIMES_amdgcn-amd-amdhsa_LLVM_ENABLE_RUNTIMES=offload
-DLLVM_RUNTIME_TARGETS=amdgcn-amd-amdhsa
```

After that, linking the OpenMP device runtime will be `-Xoffload-linker
-lomp`. I.e. no more fat binary business.

Only look at the most recent commit since this includes the two
dependencies
(fix to AMDGPUEmitPrintfBinding and the PointerToMember bug).
 [OpenMP] Fix missing type getter for SFINAE helper 2025-01-11T01:35:41+00:00 Joseph Huber huberjn@outlook.com 2025-01-11T01:35:10+00:00 74d5373f49fab0943a45a9214dc7a134f9f112f1 Summary: This didn't get the type, which made using this always return false. 
Summary:
This didn't get the type, which made using this always return false.
[OpenMP] Use __builtin_bit_cast instead of UB type punning (#122325) 2025-01-09T19:59:21+00:00 Joseph Huber huberjn@outlook.com 2025-01-09T19:59:21+00:00 f53cb84df6b80458cb4d5ab7398a590356a3a952 Summary: Use a normal bitcast, remove from the shared utils since it's not available in GCC 7.4 
Summary:
Use a normal bitcast, remove from the shared utils since it's not
available in
GCC 7.4
 [OpenMP] Update atomic helpers to just use headers (#122185) 2025-01-09T19:57:39+00:00 Joseph Huber huberjn@outlook.com 2025-01-09T19:57:39+00:00 b57c0bac81fe4f5c85c6554ca574cf2d5648e0c5 Summary: Previously we had some indirection here, this patch updates these utilities to just be normal template functions. We use SFINAE to manage the special case handling for floats. Also this strips address spaces so it can be used more generally. 
Summary:
Previously we had some indirection here, this patch updates these
utilities to just be normal template functions. We use SFINAE to manage
the special case handling for floats. Also this strips address spaces so
it can be used more generally.
 [Offload][NFC] Reorganize `utils::` and make Device/Host/Shared clearer (#100280) 2024-09-05T20:36:26+00:00 Johannes Doerfert johannes@jdoerfert.de 2024-09-05T20:36:26+00:00 08533a3ee8f3a09a59cf6ac3be59198b26b7f739 We had three `utils::` namespaces, all with different "meaning" (host, device, hsa_utils). We should, when we can, keep "include/Shared" accessible from host and device, thus RefCountTy has been moved to a separate header. `hsa_utils` was introduced to make `utils::` less overloaded. And common functionality was de-duplicated, e.g., `utils::advance` and `utils::advanceVoidPtr` -> `utils:advancePtr`. Type punning now checks for the size of the result to make sure it matches the source type. No functional change was intended. 
We had three `utils::` namespaces, all with different "meaning" (host,
device, hsa_utils). We should, when we can, keep "include/Shared"
accessible from host and device, thus RefCountTy has been moved to a
separate header. `hsa_utils` was introduced to make `utils::` less
overloaded. And common functionality was de-duplicated, e.g.,
`utils::advance` and `utils::advanceVoidPtr` -> `utils:advancePtr`. Type
punning now checks for the size of the result to make sure it matches
the source type.

No functional change was intended.