76333 Коммитов
Автор SHA1 Сообщение Дата
Edgar 4c4a7cb696 fix cmake logic to handle old and new include dirs
Starting from rocm 5.2 there is a reorganization of the
include directories. This pr allows to compile
rccl on both the old and the new directory layout.
This solution is using find_package() for identifying correct
settings for rocm_smi starting from rocm-5.2, and the original (manual)
settings for all previous releases.

Tested with rocm-5.2, 5.1.1, 5.0.2, and 4.5.2.
2022-04-28 14:33:46 -04:00
Saleel Kudchadker 29752a2bbc SWDEV-334150 - Force callback to cycle commands
Enqueue a handler callback for hipEventRecords(aka marker_ts_) for every
64 submits, This recycles the memory if we dont end up calling
synchronize for the longest time.

Change-Id: I3d39fe76d52a5d81387927edd85b5663b563682c


[ROCm/clr commit: fa76f03654]
2022-04-28 12:30:23 -04:00
Saleel Kudchadker fa76f03654 SWDEV-334150 - Force callback to cycle commands
Enqueue a handler callback for hipEventRecords(aka marker_ts_) for every
64 submits, This recycles the memory if we dont end up calling
synchronize for the longest time.

Change-Id: I3d39fe76d52a5d81387927edd85b5663b563682c
2022-04-28 12:30:23 -04:00
German Andryeyev d5bc650de9 SWDEV-307184 - Fix a regression from dafc64ea
Disable hostcall buffer in OCL for now. COv5 can add hostcallbuffer
metadata for unknown reason. OCL may fail the buffer allocation
and kernel launch.

Change-Id: I34a6a45bac86c57422b764c0d69760c96920d6c5


[ROCm/clr commit: 934149ff0a]
2022-04-28 11:57:48 -04:00
German Andryeyev 934149ff0a SWDEV-307184 - Fix a regression from 15f1632d
Disable hostcall buffer in OCL for now. COv5 can add hostcallbuffer
metadata for unknown reason. OCL may fail the buffer allocation
and kernel launch.

Change-Id: I34a6a45bac86c57422b764c0d69760c96920d6c5
2022-04-28 11:57:48 -04:00
Satyanvesh Dittakavi a6f1bde5d7 SWDEV-330667 - SWDEV-333599 - Fix hipHostRegister and hipMallocPitch APIs
- hipHostRegister should return hipErrorInvalidValue when size is 0 or
size doesn't match the actual allocated size
- hipMemAllocPitch should return hipErrorInvalidValue when width or height
is 0 and elementSizeBytes should be 4, 8 or 16

Change-Id: I43c9fe4ac1ef0a6ac2f953908d301508a7bc5aa3


[ROCm/clr commit: 06203c251e]
2022-04-28 13:07:08 +00:00
Satyanvesh Dittakavi 06203c251e SWDEV-330667 - SWDEV-333599 - Fix hipHostRegister and hipMallocPitch APIs
- hipHostRegister should return hipErrorInvalidValue when size is 0 or
size doesn't match the actual allocated size
- hipMemAllocPitch should return hipErrorInvalidValue when width or height
is 0 and elementSizeBytes should be 4, 8 or 16

Change-Id: I43c9fe4ac1ef0a6ac2f953908d301508a7bc5aa3
2022-04-28 13:07:08 +00:00
Chauncey Hui 3519b49cca SWDEV-2 - Change OpenCL version number from 3457 to 3458
[ROCm/clr commit: 75bc2d75a0]
2022-04-28 03:00:03 -04:00
Chauncey Hui 75bc2d75a0 SWDEV-2 - Change OpenCL version number from 3457 to 3458 2022-04-28 03:00:03 -04:00
Christophe Paquot c2b3dd7dc5 SWDEV-322620 - Virtual Memory Management
Implemented hipMemMap and hipMemUnmap

Change-Id: Idc75e3964d88e375e445733557cfa8e421688f1f


[ROCm/clr commit: 6d73282f25]
2022-04-28 02:14:14 -04:00
Christophe Paquot 6d73282f25 SWDEV-322620 - Virtual Memory Management
Implemented hipMemMap and hipMemUnmap

Change-Id: Idc75e3964d88e375e445733557cfa8e421688f1f
2022-04-28 02:14:14 -04:00
haoyuan2 6ed700cca8 SWDEV-333187 - Support IPC in HIP runtime on Windows
enable shared memory based IPC

Change-Id: Ic1ad8d426863c2137747a3d468fe02d073f5d248


[ROCm/clr commit: 424c4ce5c4]
2022-04-27 23:05:56 -04:00
haoyuan2 424c4ce5c4 SWDEV-333187 - Support IPC in HIP runtime on Windows
enable shared memory based IPC

Change-Id: Ic1ad8d426863c2137747a3d468fe02d073f5d248
2022-04-27 23:05:56 -04:00
gilbertlee-amd c6804778d1 [TransferBench] Syncing with TransferBench v1.02 (#541)
[ROCm/rccl commit: 685bcea127]
2022-04-27 20:43:24 -06:00
gilbertlee-amd 685bcea127 [TransferBench] Syncing with TransferBench v1.02 (#541) 2022-04-27 20:43:24 -06:00
Jonathan Kim 658b053943 Bypass HDP flush during SDMA copies on A+A GPU-CPU xGMI connections
Host to device SDMA copies do not require an HDP cache flush when
connected by xGMI since data copies over the data fabric and not HDP.

Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Sean Keely <sean.keely@amd.com>
Change-Id: I78d73a47edcc1a9c0ba59f33cf91485f13f1c45b
2022-04-27 21:45:26 -04:00
Jonathan Kim 495a3f233f Bypass HDP flush during SDMA copies on A+A GPU-CPU xGMI connections
Host to device SDMA copies do not require an HDP cache flush when
connected by xGMI since data copies over the data fabric and not HDP.

Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Sean Keely <sean.keely@amd.com>
Change-Id: I78d73a47edcc1a9c0ba59f33cf91485f13f1c45b


[ROCm/ROCR-Runtime commit: 658b053943]
2022-04-27 21:45:26 -04:00
Kiumars Sabeti 4d99f8d8e5 SWDEV-320429: wrapping the comma-containing names in the .csv in double quotes at the time the .csv is generated
Change-Id: I62f94a1cf8895eb324080f8aacac3f13c02d7050
2022-04-27 20:45:51 -04:00
Kiumars Sabeti 08c1267bbb SWDEV-320429: wrapping the comma-containing names in the .csv in double quotes at the time the .csv is generated
Change-Id: I62f94a1cf8895eb324080f8aacac3f13c02d7050


[ROCm/rocprofiler commit: 4d99f8d8e5]
2022-04-27 20:45:51 -04:00
ROCm CI Service Account 2803edb32a SWDEV-327563 - Windows: Enable most of the memory tests skipped before (#2643)
Change-Id: Ibdaf22d139f4d2772c152b2f4455676a7adbc6bd


[ROCm/hip-tests commit: 0c53bda553]
2022-04-27 15:38:07 -07:00
ROCm CI Service Account 0c53bda553 SWDEV-327563 - Windows: Enable most of the memory tests skipped before (#2643)
Change-Id: Ibdaf22d139f4d2772c152b2f4455676a7adbc6bd
2022-04-27 15:38:07 -07:00
ROCm CI Service Account 5c19483025 SWDEV-327563 - Windows: Enable most of the memory tests skipped before (#2643)
Change-Id: Ibdaf22d139f4d2772c152b2f4455676a7adbc6bd


[ROCm/hip commit: 84c7cc50b4]
2022-04-27 15:38:07 -07:00
ROCm CI Service Account 84c7cc50b4 SWDEV-327563 - Windows: Enable most of the memory tests skipped before (#2643)
Change-Id: Ibdaf22d139f4d2772c152b2f4455676a7adbc6bd
2022-04-27 15:38:07 -07:00
Jonathan R. Madsen b3c5a6f048 perfetto mpi + mpi example (#49)
[ROCm/rocprofiler-systems commit: 6b7b6e46cf]
2022-04-27 16:58:45 -05:00
Jonathan R. Madsen 6b7b6e46cf perfetto mpi + mpi example (#49) 2022-04-27 16:58:45 -05:00
Jonathan R. Madsen 2bb6fd0cfb Misc updates (#48)
- reworked `add_critical_trace`
- `get_use_thread_sampling` / `"OMNITRACE_USE_THREAD_SAMPLING"` option
- `get_cpu_cid_stack_lock`
- reworked finalization messaging
- significant updates to pthread_gotcha
  - shutdown stability
  - `"start_thread"` entries
- `rocm_smi` stability 
- roctracer_callbacks add critical trace entries on the callback thread
- reworked CPU CID initialization
- thread_sampler stability

[ROCm/rocprofiler-systems commit: 9b25d4b3b5]
2022-04-27 16:56:38 -05:00
Jonathan R. Madsen 9b25d4b3b5 Misc updates (#48)
- reworked `add_critical_trace`
- `get_use_thread_sampling` / `"OMNITRACE_USE_THREAD_SAMPLING"` option
- `get_cpu_cid_stack_lock`
- reworked finalization messaging
- significant updates to pthread_gotcha
  - shutdown stability
  - `"start_thread"` entries
- `rocm_smi` stability 
- roctracer_callbacks add critical trace entries on the callback thread
- reworked CPU CID initialization
- thread_sampler stability
2022-04-27 16:56:38 -05:00
Alex Sierra b9e8bc1f52 kfdtest: add log message to KFDExceptionTest
[WHY]
These tests force HW exceptions in the GPU driver. Some of these
exceptions might print page fault error messages at kernel level.
These are expected errors due to the nature of the tests, but still
might cause confusion to users.

[HOW]
Add log message to warn the user about these kernel error messages.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Change-Id: I8eef87b83939e37230da0c374c2f77d2d484baa9
2022-04-27 15:38:11 -05:00
Alex Sierra 05952910b1 kfdtest: add log message to KFDExceptionTest
[WHY]
These tests force HW exceptions in the GPU driver. Some of these
exceptions might print page fault error messages at kernel level.
These are expected errors due to the nature of the tests, but still
might cause confusion to users.

[HOW]
Add log message to warn the user about these kernel error messages.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Change-Id: I8eef87b83939e37230da0c374c2f77d2d484baa9


[ROCm/ROCR-Runtime commit: b9e8bc1f52]
2022-04-27 15:38:11 -05:00
Bill(Shuzhou) Liu 2cd7f66154 Update RDC document
Update README.md to refer to document portal.

Change-Id: I427122751fec5a27936b345a3ac76c96478be164
2022-04-27 14:38:48 -04:00
Bill(Shuzhou) Liu c6a69f8e59 Update RDC document
Update README.md to refer to document portal.

Change-Id: I427122751fec5a27936b345a3ac76c96478be164


[ROCm/rdc commit: 2cd7f66154]
2022-04-27 14:38:48 -04:00
Laurent Morichetti 6eb1d34cda Fix assertions
Replace EXC_ABORT() checks with assertions.

Rewrite the exception class to use std::runtime_error (as it
already handles the std::string/char* message argument).

Change-Id: I48e31924f3aea1328e6562ab6bb06ec373fd5d5e


[ROCm/roctracer commit: 5963363484]
2022-04-27 11:24:26 -07:00
Laurent Morichetti 5963363484 Fix assertions
Replace EXC_ABORT() checks with assertions.

Rewrite the exception class to use std::runtime_error (as it
already handles the std::string/char* message argument).

Change-Id: I48e31924f3aea1328e6562ab6bb06ec373fd5d5e
2022-04-27 11:24:26 -07:00
haoyuan2 558467643d SWDEV-333187 - Support IPC in HIP runtime on Windows
enable shared memory based IPC

Change-Id: I2657ef4713240b6f82a34f416d53f222630263ca


[ROCm/clr commit: 59b1ae0a79]
2022-04-27 11:14:59 -07:00
haoyuan2 59b1ae0a79 SWDEV-333187 - Support IPC in HIP runtime on Windows
enable shared memory based IPC

Change-Id: I2657ef4713240b6f82a34f416d53f222630263ca
2022-04-27 11:14:59 -07:00
Laurent Morichetti 576554dcea Fix a SEGV when running --roctx-trace
There's a typo in RegisterApiCallback, roctx::cb_table.Get should be
roctx::cb_table.Set.

Change-Id: I47ec8ac666f783ff4e03f35d13e375e645899900


[ROCm/roctracer commit: 0d7d56eea5]
2022-04-27 12:14:32 -04:00
Laurent Morichetti 0d7d56eea5 Fix a SEGV when running --roctx-trace
There's a typo in RegisterApiCallback, roctx::cb_table.Get should be
roctx::cb_table.Set.

Change-Id: I47ec8ac666f783ff4e03f35d13e375e645899900
2022-04-27 12:14:32 -04:00
Julia Jiang 1d74dfe1d5 SWDEV-334574 - Rename _bkendDevice in VDI
Change-Id: I1c04dad226e08f02bca11fa0d1981fafa7ea2d2a


[ROCm/clr commit: b7c7917256]
2022-04-27 11:21:24 -04:00
Julia Jiang b7c7917256 SWDEV-334574 - Rename _bkendDevice in VDI
Change-Id: I1c04dad226e08f02bca11fa0d1981fafa7ea2d2a
2022-04-27 11:21:24 -04:00
Chauncey Hui 21dcdac493 SWDEV-2 - Change OpenCL version number from 3456 to 3457
[ROCm/clr commit: 77b5ca79bb]
2022-04-27 03:00:11 -04:00
Chauncey Hui 77b5ca79bb SWDEV-2 - Change OpenCL version number from 3456 to 3457 2022-04-27 03:00:11 -04:00
Ranjith Ramakrishnan 140d2d4bc2 Merge "Populate roctracer.h wrapper file with orginal file contents as dead code" into amd-staging
[ROCm/roctracer commit: 7f05496a87]
2022-04-27 02:20:26 -04:00
Ranjith Ramakrishnan 7f05496a87 Merge "Populate roctracer.h wrapper file with orginal file contents as dead code" into amd-staging 2022-04-27 02:20:26 -04:00
Jonathan R. Madsen d45e84b116 GOTCHA + Kokkos + tasking + more (#47)
* GOTCHA + Kokkos + tasking + more

- update gotcha with fix for dlsym(RTLD_NEXT, ...)
- support for standalone KOKKOS_PROFILE_LIBRARY
- remove extra flags for omnitrace-user
- roctracer and critical_trace namespaces in tasking
- generic tasking functions, e.g. join(), shutdown(), etc.
- omnitrace_init_tooling_hidden in api.hpp
- ompt.cpp uses OMNITRACE_USE_OMPT
- kokkosp uses user_region instead of omnitrace component
- re-enable recycling thread ids
- more generic _{push,pop}_perfetto functors
- fix for thread_data::instance(construct_on_init, ...)
- fix for omnitrace-headers interface target
- omnitrace_watch_for_change

[ROCm/rocprofiler-systems commit: 29220cba58]
2022-04-26 22:08:51 -05:00
Jonathan R. Madsen 29220cba58 GOTCHA + Kokkos + tasking + more (#47)
* GOTCHA + Kokkos + tasking + more

- update gotcha with fix for dlsym(RTLD_NEXT, ...)
- support for standalone KOKKOS_PROFILE_LIBRARY
- remove extra flags for omnitrace-user
- roctracer and critical_trace namespaces in tasking
- generic tasking functions, e.g. join(), shutdown(), etc.
- omnitrace_init_tooling_hidden in api.hpp
- ompt.cpp uses OMNITRACE_USE_OMPT
- kokkosp uses user_region instead of omnitrace component
- re-enable recycling thread ids
- more generic _{push,pop}_perfetto functors
- fix for thread_data::instance(construct_on_init, ...)
- fix for omnitrace-headers interface target
- omnitrace_watch_for_change
2022-04-26 22:08:51 -05:00
pghafari 0ddbd07e0c SWDEV-327960, SWDEV-316901 - Adding Graph checks
hipGraphExecMemcpyNodeSetParamsToSymbol, fix for functional cases
failing as well

Change-Id: I9f263778f6b24e3ec35a78231a17e01c0d44f23a


[ROCm/clr commit: 63f23ec2b0]
2022-04-26 21:04:32 -04:00
pghafari 63f23ec2b0 SWDEV-327960, SWDEV-316901 - Adding Graph checks
hipGraphExecMemcpyNodeSetParamsToSymbol, fix for functional cases
failing as well

Change-Id: I9f263778f6b24e3ec35a78231a17e01c0d44f23a
2022-04-26 21:04:32 -04:00
Jenkins 16ff90454e SWDEV-299127 - Merge 'develop' into 'amd-staging'
Change-Id: I54aa4822132349d32cf912601adead0812c23230


[ROCm/hip commit: f9764efdfd]
2022-04-26 19:11:28 -04:00
Jenkins f9764efdfd SWDEV-299127 - Merge 'develop' into 'amd-staging'
Change-Id: I54aa4822132349d32cf912601adead0812c23230
2022-04-26 19:11:28 -04:00
Wenkai Du 95b30d9762 topo_expl: fix build and add tuning support (#539)
[ROCm/rccl commit: 063da25563]
2022-04-26 15:40:07 -07:00