SWDEV-145570 - Adding back the lazy kernel changes because the OOM issue is because of KFD/RocR.
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_platform.cpp#53 edit
+ Both Driver API and RT API are supported and synced with each other
+ Update *.md docs and hipify-perl accordingly
+ Add new conversion type "virtual_memory", introduced in Dirver API
[Background] it was found that if lazy linking used for a library that calls hipExtLaunchMultiKernelMultiDevice API then this API can get the wrong program_state object for looking up device kernels leading to a "No device code available" error in this API.
To fix this issue, the API was refactored to be inline and get and pass the correct program_state to an internal hip API to request a multi-device kernel launch.
SWDEV-214490 - Update HIP RT for texture3D in HIP/PAL on Windows
-Update ihipBindTexture
http://ocltc.amd.com/reviews/r/18333/
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#89 edit
... //depot/stg/opencl/drivers/opencl/api/hip/hip_texture.cpp#28 edit
+ Update *.md docs and hipify-perl accordingly
[Reason]
Starting with CUDA 10.1 all error codes are merged between Driver and RT APIs
[ToDo]
Do the same merge in HIP API as there is no need in distinguishing return codes by API
SWDEV-214490 - Update HIP RT for texture3D in HIP/PAL on Windows
- Update function hipMemcpy3D for Texture Array
- Add hipArrayCubemap support in hipMalloc3DArray
http://ocltc.amd.com/reviews/r/18328/
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#88 edit
SWDEV-213526 - pytorch tests fail with hipErrorOutofMemory
There's a bug in ROCr when loading a lot of kernels and not syncing
So for now, if an allocation fails, sync devices and retry before
returning hipErrorOutOfMemory error.
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#86 edit
SWDEV-212749:
o Recent changes to “add support for extended launch” require hip_runtime.h to be include in hip_ext.h
o Order in which external applications include hip_hcc.h/hip_runtime.h causes compilation failure
This will fix issue #1621. It also adds tests for is_callable with c++11, c++14, and c++17.
The fallback implementation was completely broken so I rewrote it so it pass the tests as well. This should be used instead of PR #1631.
* Updated hipEnvVarDriver to work with Windows
* Cleaned up a bit of code
* Fixed a part where putenv was used for both win and linux
* Defines moved to test_common.h and cleaned up code
* Cleaned up some macro defines and used const char instead
* Got rid of some excess commenting
* directory paths are unconditional
* Cleaned some duplicate code, and variables are now declared and defined together
* Add support for extended launch syntax.
* Add unit test.
* Fix typo
* hipExtLaunchKernelGGL lives in hip_ext.h
Change-Id: Ice32dab0d43475fda65c6a910c11416871a8f2ff
* [dtest] remove redundant include from hipModuleGetGlobal dtest
SWDEV-212440 - [HIP] Memory access fault observed on Pytorch while running performance tests with Microbenchmarking script
We need to loop through all the default stream to sync them in case
the app call hipFree on a different current stream and another current stream
is using the memory.
Affected files ...
... //depot/stg/opencl/drivers/opencl/api/hip/hip_memory.cpp#85 edit
foreman
Evgeny Mankov
Maneesh Gupta
Rahul Garg
Aryan Salmanpour
Jenkins
Siu Chi Chan
Alex Voicu
Jeff Daily
Anusha Godavarthy Surya
Shyam Saladi
ansurya
satyanveshd
Paul Fultz II
Sarbojit2019
Sameer Sahasrabuddhe
mhbliao
Nick Curtis
jiabaxie
kjayapra-amd