29efd8eccd
VMM: Fix flags when allocating memory handle
...
When allocating a memory handle, the NoAddress thunk flag should be set
so that this allocation does not have a virtual address range.
Also, skip mapping the memory when allocating a memory handle
Change-Id: I1c168bc00ddbc158d447197c4dc25f96bad02b19
2024-01-17 10:24:58 -05:00
7c35d797d9
VMM: Fix flags when allocating memory handle
...
When allocating a memory handle, the NoAddress thunk flag should be set
so that this allocation does not have a virtual address range.
Also, skip mapping the memory when allocating a memory handle
Change-Id: I1c168bc00ddbc158d447197c4dc25f96bad02b19
[ROCm/ROCR-Runtime commit: 29efd8eccd
2024-01-17 10:24:58 -05:00
2f97049da5
VMM: Default access should be none
...
After a memory handle is created. hsa_amd_vmem_get_access should return
HSA_ACCESS_PERMISSION_NONE insread of reporting the allocation as
invalid.
Change-Id: I1a09d15c220d48497d09c89059493e538f82aeb9
2024-01-17 10:24:51 -05:00
a8664a7471
VMM: Default access should be none
...
After a memory handle is created. hsa_amd_vmem_get_access should return
HSA_ACCESS_PERMISSION_NONE insread of reporting the allocation as
invalid.
Change-Id: I1a09d15c220d48497d09c89059493e538f82aeb9
[ROCm/ROCR-Runtime commit: 2f97049da5
2024-01-17 10:24:51 -05:00
8b85f9e668
VMM: Fix access for multi-GPU
...
When using multi-GPU for each BO, a new dmabuf_fd needs to be imported
into libdrm.
Change-Id: Iaa2415c8f655a1ce8e92b0878517a11ff014a1d5
2024-01-17 10:24:35 -05:00
01ae507bf2
VMM: Fix access for multi-GPU
...
When using multi-GPU for each BO, a new dmabuf_fd needs to be imported
into libdrm.
Change-Id: Iaa2415c8f655a1ce8e92b0878517a11ff014a1d5
[ROCm/ROCR-Runtime commit: 8b85f9e668
2024-01-17 10:24:35 -05:00
b32ac029a8
SWDEV-429802 - change address sanitizer variable name to ENABLE_ADDRESS_SANITIZER
...
This is to bring scripts in compute/ec/prototype in sync which uses this
particular variable to enable address sanitizer across projects.
Change-Id: Ia620e4fd86338554ee3dd57c5690edb919b6e5a5
[ROCm/hip-tests commit: 508991225f
2024-01-17 10:09:51 -05:00
508991225f
SWDEV-429802 - change address sanitizer variable name to ENABLE_ADDRESS_SANITIZER
...
This is to bring scripts in compute/ec/prototype in sync which uses this
particular variable to enable address sanitizer across projects.
Change-Id: Ia620e4fd86338554ee3dd57c5690edb919b6e5a5
2024-01-17 10:09:51 -05:00
1b05247a03
SWDEV-359287 - uniform_work_group_size is interger type is 1 for true and 0 for false. Lighting kernel's init sets uniform work group flag and hence not need to set it from options.
...
Change-Id: I8d3a9a6ed1c3bddaaa86f7cdc242a45e6a1d09f6
[ROCm/clr commit: ea25e65834
2024-01-17 04:42:52 -05:00
ea25e65834
SWDEV-359287 - uniform_work_group_size is interger type is 1 for true and 0 for false. Lighting kernel's init sets uniform work group flag and hence not need to set it from options.
...
Change-Id: I8d3a9a6ed1c3bddaaa86f7cdc242a45e6a1d09f6
2024-01-17 04:42:52 -05:00
e9cb492880
SWDEV-412457 - add algorithm header in file
...
Change-Id: I18549f612be9e0b51d4ef92c600f1ce72ab58759
[ROCm/hip-tests commit: 252fca5aa8
2024-01-17 09:20:34 +00:00
252fca5aa8
SWDEV-412457 - add algorithm header in file
...
Change-Id: I18549f612be9e0b51d4ef92c600f1ce72ab58759
2024-01-17 09:20:34 +00:00
3e038d871c
cloning doxygen-awesome-css to external ( #389 )
...
[ROCm/rocprofiler-sdk commit: b55cea0e98
2024-01-17 02:44:52 -06:00
b55cea0e98
cloning doxygen-awesome-css to external ( #389 )
2024-01-17 02:44:52 -06:00
067071f669
SWDEV-439757: Remove codeobj marker on last kernel. Fixed codeobj size limit. ATT version bump.
...
Change-Id: Ie0b9c4de6c37acaaad4ae7d8d774d234a3847b7c
2024-01-17 05:01:50 -03:00
68cfdd66ba
SWDEV-439757: Remove codeobj marker on last kernel. Fixed codeobj size limit. ATT version bump.
...
Change-Id: Ie0b9c4de6c37acaaad4ae7d8d774d234a3847b7c
[ROCm/rocprofiler commit: 067071f669
2024-01-17 05:01:50 -03:00
e8c7e6bfba
ROCTx Library ( #360 )
...
* Initial implementation of roctx library
* Update include/roctx/CMakeLists.txt
- fix installation
* Update cmake/rocprofiler_config_packaging.cmake
- add rocprofiler-sdk-roctx installer
* Update include/roctx/CMakeLists.txt
- include api_trace.h in installation
* Update include/roctx/api_trace.h
- add ROCTX_API_TABLE_VERSION_MAJOR define
- add ROCTX_API_TABLE_VERSION_STEP define
* Update lib/roctx/roctx.cpp
- static asserts for table size and struct member offsets
* Update external/CMakeLists.txt
- move BUILD_SHARED_LIBS to top
- disable libunwind for glog
* Update lib/roctx/CMakeLists.txt
- Update {BUILD,INSTALL}_RPATH
* Relocate include/roctx to include/rocprofiler-sdk/roctx
* Relocate lib/roctx to lib/rocprofiler-sdk-roctx
- change the name of the library from libroctx to librocprofiler-sdk-roctx
* Move lib/plugins to lib/rocprofiler-sdk-tool/plugins
- also change install export group
* Update lib/rocprofiler-sdk/CMakeLists.txt
- change rocprofiler-shared-library EXPORT group (rocprofiler-sdk-library-targets -> rocprofiler-sdk-targets)
* Update cmake/rocprofiler_utilities.cmake
- change install EXPORT group
- rocprofiler-sdk-library-targets -> rocprofiler-sdk-targets
* Update CMakeLists.txt
- set PACKAGE_NAME at high level
- include(rocprofiler_config_install_roctx)
* Update cmake/rocprofiler_config_install* and cmake/Templates/*.cmake.in
- added rocprofiler_config_install_roctx.cmake for installing roctx as a package
- reorganization of existing cmake/Templates/*-config.cmake.in files
- created new config.cmake.in and build-config.cmake.in for rocprofiler-sdk-roctx
* Relocate include/rocprofiler-sdk/roctx to include/rocprofiler-sdk-roctx
* Update rocprofiler_config_install_roctx.cmake
* Update lib/rocprofiler-sdk-roctx/roctx.cpp
- update include paths
* Update lib/rocprofiler-sdk-roctx/CMakeLists.txt
- change target name to have rocprofiler-sdk- prefix
- interface target_include_directories
- define export symbol
* source formatting (clang-format v11) (#361 )
Co-authored-by: jrmadsen <jrmadsen@users.noreply.github.com >
* Update include/rocprofiler-sdk/fwd.h
- fix doxygen markup for ROCPROFILER_STATUS_ERROR_CONTEXT_ERROR
* Update modulefile and setup-env.sh
* Update cmake/Templates/rocprofiler-sdk/config.cmake.in
- fix inclusion of rocprofiler-sdk-targets.cmake
* Update include/rocprofiler-sdk-roctx
- add types.h for typedefs
- add doxygen comments for roctx.h
- add roctxGetThreadId function
- roctxProfilerStart and roctxProfilerStop accept thread ID param
* Update lib/rocprofiler-sdk-roctx/roctx.cpp
- hsa_agent_t* -> hsa_agent_s*
* Update lib/rocprofiler-sdk-roctx/roctx.cpp
- support for roctxGetThreadId
- update signatures of roctxProfilerPause and roctxProfilerResume
* Update lib/rocprofiler-sdk-roctx/roctx.cpp
- Initialize logging with ROCTX_LOG_LEVEL
* Update include/rocprofiler-sdk-roctx/roctx.h
- remove ROCTX_NONNULL for ihipStream_t parameter in roctxNameHipStream because default stream is a nullptr
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
[ROCm/rocprofiler-sdk commit: 1edd4891b2
2024-01-17 01:27:41 -06:00
1edd4891b2
ROCTx Library ( #360 )
...
* Initial implementation of roctx library
* Update include/roctx/CMakeLists.txt
- fix installation
* Update cmake/rocprofiler_config_packaging.cmake
- add rocprofiler-sdk-roctx installer
* Update include/roctx/CMakeLists.txt
- include api_trace.h in installation
* Update include/roctx/api_trace.h
- add ROCTX_API_TABLE_VERSION_MAJOR define
- add ROCTX_API_TABLE_VERSION_STEP define
* Update lib/roctx/roctx.cpp
- static asserts for table size and struct member offsets
* Update external/CMakeLists.txt
- move BUILD_SHARED_LIBS to top
- disable libunwind for glog
* Update lib/roctx/CMakeLists.txt
- Update {BUILD,INSTALL}_RPATH
* Relocate include/roctx to include/rocprofiler-sdk/roctx
* Relocate lib/roctx to lib/rocprofiler-sdk-roctx
- change the name of the library from libroctx to librocprofiler-sdk-roctx
* Move lib/plugins to lib/rocprofiler-sdk-tool/plugins
- also change install export group
* Update lib/rocprofiler-sdk/CMakeLists.txt
- change rocprofiler-shared-library EXPORT group (rocprofiler-sdk-library-targets -> rocprofiler-sdk-targets)
* Update cmake/rocprofiler_utilities.cmake
- change install EXPORT group
- rocprofiler-sdk-library-targets -> rocprofiler-sdk-targets
* Update CMakeLists.txt
- set PACKAGE_NAME at high level
- include(rocprofiler_config_install_roctx)
* Update cmake/rocprofiler_config_install* and cmake/Templates/*.cmake.in
- added rocprofiler_config_install_roctx.cmake for installing roctx as a package
- reorganization of existing cmake/Templates/*-config.cmake.in files
- created new config.cmake.in and build-config.cmake.in for rocprofiler-sdk-roctx
* Relocate include/rocprofiler-sdk/roctx to include/rocprofiler-sdk-roctx
* Update rocprofiler_config_install_roctx.cmake
* Update lib/rocprofiler-sdk-roctx/roctx.cpp
- update include paths
* Update lib/rocprofiler-sdk-roctx/CMakeLists.txt
- change target name to have rocprofiler-sdk- prefix
- interface target_include_directories
- define export symbol
* source formatting (clang-format v11) (#361 )
Co-authored-by: jrmadsen <jrmadsen@users.noreply.github.com >
* Update include/rocprofiler-sdk/fwd.h
- fix doxygen markup for ROCPROFILER_STATUS_ERROR_CONTEXT_ERROR
* Update modulefile and setup-env.sh
* Update cmake/Templates/rocprofiler-sdk/config.cmake.in
- fix inclusion of rocprofiler-sdk-targets.cmake
* Update include/rocprofiler-sdk-roctx
- add types.h for typedefs
- add doxygen comments for roctx.h
- add roctxGetThreadId function
- roctxProfilerStart and roctxProfilerStop accept thread ID param
* Update lib/rocprofiler-sdk-roctx/roctx.cpp
- hsa_agent_t* -> hsa_agent_s*
* Update lib/rocprofiler-sdk-roctx/roctx.cpp
- support for roctxGetThreadId
- update signatures of roctxProfilerPause and roctxProfilerResume
* Update lib/rocprofiler-sdk-roctx/roctx.cpp
- Initialize logging with ROCTX_LOG_LEVEL
* Update include/rocprofiler-sdk-roctx/roctx.h
- remove ROCTX_NONNULL for ihipStream_t parameter in roctxNameHipStream because default stream is a nullptr
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
2024-01-17 01:27:41 -06:00
12bc6c93d7
Cleanup + logging env variable ( #387 )
...
* [CP] Update tests/common/serialization.hpp
- remove duplication in rocprofiler_callback_tracing_code_object_load_data_t
* [CP] Update lib/rocprofiler-sdk/tests
- create common.hpp
- update registration.cpp to use common.hpp
* [CP] Add lib/common/logging.{hpp,cpp}
- generic init_logging function
* [CP] Update lib/rocprofiler-sdk/hsa/async_copy.cpp
- remove excess logging
* [CP] Update lib/rocprofiler-sdk/registration.cpp
- use common::init_logging(...)
- enforce ROCPROFILER_REGISTER_FORCE_LOAD in rocprofiler_force_configure
- logging updates in rocprofiler_set_api_table
* Update include/rocprofiler-sdk/buffer_tracing.h
- rocprofiler_buffer_tracing_marker_record_t -> rocprofiler_buffer_tracing_marker_api_record_t
* Update lib/common/utility.hpp
- remove active_capacity_gate
* Update lib/rocprofiler-sdk/tests/common.hpp
- fix get_{callback,buffer}_tracing_names()
* Update lib/rocprofiler-sdk/counters/xml/{basic,derived}_counters.xml
- add entries for gfx1102
[ROCm/rocprofiler-sdk commit: dc8b8aa448
2024-01-17 00:28:20 -06:00
dc8b8aa448
Cleanup + logging env variable ( #387 )
...
* [CP] Update tests/common/serialization.hpp
- remove duplication in rocprofiler_callback_tracing_code_object_load_data_t
* [CP] Update lib/rocprofiler-sdk/tests
- create common.hpp
- update registration.cpp to use common.hpp
* [CP] Add lib/common/logging.{hpp,cpp}
- generic init_logging function
* [CP] Update lib/rocprofiler-sdk/hsa/async_copy.cpp
- remove excess logging
* [CP] Update lib/rocprofiler-sdk/registration.cpp
- use common::init_logging(...)
- enforce ROCPROFILER_REGISTER_FORCE_LOAD in rocprofiler_force_configure
- logging updates in rocprofiler_set_api_table
* Update include/rocprofiler-sdk/buffer_tracing.h
- rocprofiler_buffer_tracing_marker_record_t -> rocprofiler_buffer_tracing_marker_api_record_t
* Update lib/common/utility.hpp
- remove active_capacity_gate
* Update lib/rocprofiler-sdk/tests/common.hpp
- fix get_{callback,buffer}_tracing_names()
* Update lib/rocprofiler-sdk/counters/xml/{basic,derived}_counters.xml
- add entries for gfx1102
2024-01-17 00:28:20 -06:00
8f0ea44c09
Suppress reporting no tools were found with rocprofiler-register
...
Change-Id: If853517d40e073202d12e2a6b16fb54be5529650
2024-01-17 01:01:19 -05:00
4f7dfe87d2
Suppress reporting no tools were found with rocprofiler-register
...
Change-Id: If853517d40e073202d12e2a6b16fb54be5529650
[ROCm/ROCR-Runtime commit: 8f0ea44c09
2024-01-17 01:01:19 -05:00
871048b7b9
Add check to ensure metrics are valid on GPU Arch ( #384 )
...
* Add check to ensure metrics are valid on GPU Arch
Ensure requested metrics are valid on the GPU arch. If not valid,
error is returned during profile config init.
* source formatting (clang-format v11) (#385 )
Co-authored-by: bwelton <bwelton@users.noreply.github.com >
* Update metrics.cpp
* source formatting (clang-format v11) (#386 )
Co-authored-by: bwelton <bwelton@users.noreply.github.com >
* Update metrics.cpp
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: bwelton <bwelton@users.noreply.github.com >
[ROCm/rocprofiler-sdk commit: 0952308c4a
2024-01-16 21:47:45 -06:00
0952308c4a
Add check to ensure metrics are valid on GPU Arch ( #384 )
...
* Add check to ensure metrics are valid on GPU Arch
Ensure requested metrics are valid on the GPU arch. If not valid,
error is returned during profile config init.
* source formatting (clang-format v11) (#385 )
Co-authored-by: bwelton <bwelton@users.noreply.github.com >
* Update metrics.cpp
* source formatting (clang-format v11) (#386 )
Co-authored-by: bwelton <bwelton@users.noreply.github.com >
* Update metrics.cpp
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: bwelton <bwelton@users.noreply.github.com >
2024-01-16 21:47:45 -06:00
0accd17b6e
HSA_USE_SVM to override SVMAPISupported node prop
...
When HSA_USE_SVM is 0, thunk uses non-SVM path, but upper layers still
use SVM path. That is not as expected.
Suggested-by: Lang Yu <Lang.Yu@amd.com >
Change-Id: I1ae0b4faa2f8af5ec69a81cfeb7661bd47d739d4
2024-01-16 22:44:38 -05:00
68522c65a9
HSA_USE_SVM to override SVMAPISupported node prop
...
When HSA_USE_SVM is 0, thunk uses non-SVM path, but upper layers still
use SVM path. That is not as expected.
Suggested-by: Lang Yu <Lang.Yu@amd.com >
Change-Id: I1ae0b4faa2f8af5ec69a81cfeb7661bd47d739d4
[ROCm/ROCR-Runtime commit: 0accd17b6e
2024-01-16 22:44:38 -05:00
e20f41df62
Enable IPC DMA buf
...
Set HSA_ENABLE_IPC_MODE_LEGACY off (i.e. use DMA bufs implementation
by default).
Change-Id: I7b1c6cb7d19310adf6f0bfe060736f4adbf7adc2
2024-01-16 22:43:27 -05:00
7ac11c41af
Enable IPC DMA buf
...
Set HSA_ENABLE_IPC_MODE_LEGACY off (i.e. use DMA bufs implementation
by default).
Change-Id: I7b1c6cb7d19310adf6f0bfe060736f4adbf7adc2
[ROCm/ROCR-Runtime commit: e20f41df62
2024-01-16 22:43:27 -05:00
5dfebdbca9
Change IPC implementation to use DMA Bufs
...
As the KFD IPC IOCTLs will not be upstreamed, change runtime
implementation to use DMA bufs.
DMA buf fds will be passed over abstract unix domain sockets.
The exporter spins a thread that creates a socket server.
The importer connects to the server to fetch the fd.
libDRM will be required to do a manual import and GPU map for
memory that is not already imported and mapped.
For now, use the legacy IPC implementation by default as a
follow on patch will disable the HSA_ENABLE_IPC_MODE_LEGACY
environment variable.
Change-Id: Ifd8469e9adfc81f8a1ea78d6010fb10b515ba1b4
2024-01-16 22:43:00 -05:00
3c63cf521b
Change IPC implementation to use DMA Bufs
...
As the KFD IPC IOCTLs will not be upstreamed, change runtime
implementation to use DMA bufs.
DMA buf fds will be passed over abstract unix domain sockets.
The exporter spins a thread that creates a socket server.
The importer connects to the server to fetch the fd.
libDRM will be required to do a manual import and GPU map for
memory that is not already imported and mapped.
For now, use the legacy IPC implementation by default as a
follow on patch will disable the HSA_ENABLE_IPC_MODE_LEGACY
environment variable.
Change-Id: Ifd8469e9adfc81f8a1ea78d6010fb10b515ba1b4
[ROCm/ROCR-Runtime commit: 5dfebdbca9
2024-01-16 22:43:00 -05:00
099db3bab7
Drop signal limiter ( #382 )
...
* Drop signal limiter
* source formatting (clang-format v11) (#383 )
Co-authored-by: bwelton <bwelton@users.noreply.github.com >
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: bwelton <bwelton@users.noreply.github.com >
[ROCm/rocprofiler-sdk commit: e0fd4c1716
2024-01-16 19:43:24 -06:00
e0fd4c1716
Drop signal limiter ( #382 )
...
* Drop signal limiter
* source formatting (clang-format v11) (#383 )
Co-authored-by: bwelton <bwelton@users.noreply.github.com >
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: bwelton <bwelton@users.noreply.github.com >
2024-01-16 19:43:24 -06:00
baaad8b45a
SWDEV-398109 - Enable smid support for CU mode for gfx10+.
...
Change-Id: Ief1e7c952c3de7e4a547bbdcc9817f51aa4d5d36
[ROCm/clr commit: fb2c100d8c
2024-01-16 20:13:23 -05:00
fb2c100d8c
SWDEV-398109 - Enable smid support for CU mode for gfx10+.
...
Change-Id: Ief1e7c952c3de7e4a547bbdcc9817f51aa4d5d36
2024-01-16 20:13:23 -05:00
0e3f668e2c
Use HybridMutex for IPC locks
...
Change-Id: I24ab4a96237612a7d32beda06cc20b25cb1f0b37
2024-01-16 21:29:39 +00:00
ba59a36d8a
Use HybridMutex for IPC locks
...
Change-Id: I24ab4a96237612a7d32beda06cc20b25cb1f0b37
[ROCm/ROCR-Runtime commit: 0e3f668e2c
2024-01-16 21:29:39 +00:00
8d3fee5095
Use HybridMutex for signal mutexes
...
Implement HybridMutex to improve latencies compared to KernelMutex when
there is contention between several threads calling hsa_signal_create
and hsa_amd_signal_async_handler.
Change-Id: If53377033e749b0050727964c9303f09b02527cc
2024-01-16 21:29:39 +00:00
cbe9337918
Use HybridMutex for signal mutexes
...
Implement HybridMutex to improve latencies compared to KernelMutex when
there is contention between several threads calling hsa_signal_create
and hsa_amd_signal_async_handler.
Change-Id: If53377033e749b0050727964c9303f09b02527cc
[ROCm/ROCR-Runtime commit: 8d3fee5095
2024-01-16 21:29:39 +00:00
3d1563ee68
Force t1_ update when profiling is enabled
...
Fixes issue where t1_ counters may not be updated when doing dispatch
profiling, causing a divide by 0.
Change-Id: I91060ac3f9fd2183d277e6e7cd810398a453a87f
2024-01-16 21:29:39 +00:00
dc9bce3b9b
Force t1_ update when profiling is enabled
...
Fixes issue where t1_ counters may not be updated when doing dispatch
profiling, causing a divide by 0.
Change-Id: I91060ac3f9fd2183d277e6e7cd810398a453a87f
[ROCm/ROCR-Runtime commit: 3d1563ee68
2024-01-16 21:29:39 +00:00
d16c6db2ee
Increase min KFD version for Virtual mem support
...
KFD had some fixes for handling of virtual memory APIs. These fixes are
included in interface version 1.15.
Change-Id: Ie701eccf6e032f9ec0a1f4e8a43718964eebddc6
2024-01-16 21:29:39 +00:00
bc492274e7
Increase min KFD version for Virtual mem support
...
KFD had some fixes for handling of virtual memory APIs. These fixes are
included in interface version 1.15.
Change-Id: Ie701eccf6e032f9ec0a1f4e8a43718964eebddc6
[ROCm/ROCR-Runtime commit: d16c6db2ee
2024-01-16 21:29:39 +00:00
4971150576
Improve endianness check
...
Update the `hsa.h` header to use the gcc / clang `__BYTE_ORDER__`
macros where available to more accurately autodetect endianness for
the target.
Change-Id: I7312f3badcba9287a30eb14882b91e2a247acc5f
2024-01-16 21:29:39 +00:00
298e6cc495
Improve endianness check
...
Update the `hsa.h` header to use the gcc / clang `__BYTE_ORDER__`
macros where available to more accurately autodetect endianness for
the target.
Change-Id: I7312f3badcba9287a30eb14882b91e2a247acc5f
[ROCm/ROCR-Runtime commit: 4971150576
2024-01-16 21:29:39 +00:00
0fe462b63e
SWDEV-441258 - remove full path for HIP LOG windows
...
Change-Id: Ibad6e9542c0cede38f5a114dcd352356ddedf019
[ROCm/clr commit: 0cff14c9e1
2024-01-16 15:26:06 -05:00
0cff14c9e1
SWDEV-441258 - remove full path for HIP LOG windows
...
Change-Id: Ibad6e9542c0cede38f5a114dcd352356ddedf019
2024-01-16 15:26:06 -05:00
37ed51c99c
SWDEV-440746 - Enable hipExtAnyOrderLaunch extension for PAL
...
Extension allows to execute the kernels without a wait barrier and L1
invalidation.
Change-Id: I96c485204303f54a0240b93134f4560673e4bd17
[ROCm/clr commit: 13c6f56ca9
2024-01-16 15:20:39 -05:00
13c6f56ca9
SWDEV-440746 - Enable hipExtAnyOrderLaunch extension for PAL
...
Extension allows to execute the kernels without a wait barrier and L1
invalidation.
Change-Id: I96c485204303f54a0240b93134f4560673e4bd17
2024-01-16 15:20:39 -05:00
9af6945c84
SWDEV-421025 - hipGraphInstantiateWithParams structs and enums
...
Change-Id: Ia5353589c44aff1dc37b41c385ce8eed6bf9f42d
[ROCm/hip commit: 72a245881c
2024-01-16 13:36:14 -05:00
72a245881c
SWDEV-421025 - hipGraphInstantiateWithParams structs and enums
...
Change-Id: Ia5353589c44aff1dc37b41c385ce8eed6bf9f42d
2024-01-16 13:36:14 -05:00
David Yat Sin
Jatin Chaudhary
Jaydeep Patel
Gopesh Bhardwaj
Giovanni LB
Jonathan R. Madsen
Jonathan R. Madsen
Benjamin Welton
Jonathan Kim
Joseph Huber
pghafari
German
Rahul Manocha