AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
64 512 Commits 3 Branches 0 Tags
bb1295bcdfd72123700cd66e631326ab38e82db8
Commit Graph

64512 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
vlaindic 03a63791b0 SWDEV-418738: Listing counters of all GPU agents 
Using `gfxip` when listing counters on the system with multi-GPUs
all belonging to the same family results in showing counters
of the first agent multiple times.
Instead, use the agent's `name` to observe counters of all agents.

Change-Id: I7b46c0670bc7c918e2ba0357fa659df9d23379d4
 2023-09-22 17:22:28 +02:00
vlaindic 581bf2cd87 SWDEV-418738: Listing counters of all GPU agents 
Using `gfxip` when listing counters on the system with multi-GPUs
all belonging to the same family results in showing counters
of the first agent multiple times.
Instead, use the agent's `name` to observe counters of all agents.

Change-Id: I7b46c0670bc7c918e2ba0357fa659df9d23379d4


[ROCm/rocprofiler commit: 03a63791b0]
 2023-09-22 17:22:28 +02:00
David Yat Sin 8e06dce573 Add extended coherence memory flag 
Add support for new flag for memory allocation that will provide
system-scope coherent atomics

Change-Id: I426d66223e8d2b570f69b4c0e61145ce9b2290d2
 2023-09-22 11:03:00 -04:00
David Yat Sin 351cbe9dc7 Add extended coherence memory flag 
Add support for new flag for memory allocation that will provide
system-scope coherent atomics

Change-Id: I426d66223e8d2b570f69b4c0e61145ce9b2290d2


[ROCm/ROCR-Runtime commit: 8e06dce573]
 2023-09-22 11:03:00 -04:00
Jatin Chaudhary f62807d411 SWDEV-421819 - Add ptrdiff_t and clock_t. 
Rework types inside new __HIP_NO_STD_DEFS__ guard

Change-Id: I2b67381c01ba497b323864afe0a8d1402e396748


[ROCm/clr commit: 49bb6a4072]
 2023-09-22 05:07:52 -04:00
Jatin Chaudhary 49bb6a4072 SWDEV-421819 - Add ptrdiff_t and clock_t. 
Rework types inside new __HIP_NO_STD_DEFS__ guard

Change-Id: I2b67381c01ba497b323864afe0a8d1402e396748
 2023-09-22 05:07:52 -04:00
pghafari f088f06be6 SWDEV-369567 - [ABI Break] updating gridY and Z max dimension check 
Change-Id: If515fda4938d3c5f455a02b4a65adaad76dc0aed


[ROCm/clr commit: 96f41fad07]
 2023-09-22 04:52:27 -04:00
pghafari 96f41fad07 SWDEV-369567 - [ABI Break] updating gridY and Z max dimension check 
Change-Id: If515fda4938d3c5f455a02b4a65adaad76dc0aed
 2023-09-22 04:52:27 -04:00
sdashmiz 0e64e42c60 SWDEV-330665 - [ABI Break] check for attribute 
- Add support for DevAttrHostRegisterSupported

Signed-off-by: sdashmiz <shadi.dashmiz@amd.com>
Change-Id: Ica8d323eda97b65ef0088aa99c84dd36cfe4edf3


[ROCm/clr commit: 8705322f10]
 2023-09-22 04:51:24 -04:00
sdashmiz 8705322f10 SWDEV-330665 - [ABI Break] check for attribute 
- Add support for DevAttrHostRegisterSupported

Signed-off-by: sdashmiz <shadi.dashmiz@amd.com>
Change-Id: Ica8d323eda97b65ef0088aa99c84dd36cfe4edf3
 2023-09-22 04:51:24 -04:00
Satyanvesh Dittakavi 9ee9f04cdd SWDEV-407358 - Address MIGraphx -Werror issue with unsafe buffer access 
Change-Id: If4189dfbf12b670af60782f09ff8cf9b9867c299


[ROCm/clr commit: 16fb3ca923]
 2023-09-22 04:50:32 -04:00
Satyanvesh Dittakavi 16fb3ca923 SWDEV-407358 - Address MIGraphx -Werror issue with unsafe buffer access 
Change-Id: If4189dfbf12b670af60782f09ff8cf9b9867c299
 2023-09-22 04:50:32 -04:00
Rakesh Roy 8d9ee90534 SWDEV-1 - Disable randomly failing tests on Nvidia 
- These tests are failing randomly on Nvidia during PSDB

Change-Id: Ifb4d085a526c1ff7d1b7af0528cc94d536b99db9


[ROCm/hip-tests commit: fa0fd5cb83]
 2023-09-22 04:29:21 -04:00
Rakesh Roy fa0fd5cb83 SWDEV-1 - Disable randomly failing tests on Nvidia 
- These tests are failing randomly on Nvidia during PSDB

Change-Id: Ifb4d085a526c1ff7d1b7af0528cc94d536b99db9
 2023-09-22 04:29:21 -04:00
Sourabh Betigeri 76bc392c3b SWDEV-396656 - Adds CG tiled_partition and coalesced_groups tests 
- Tests meta_group_rank/size() in static tile
               partitioning

Change-Id: I1a0d83224145a2356aa12b5e83b094834b53d5a4


[ROCm/hip-tests commit: 850a18bf2f]
 2023-09-22 04:49:19 +00:00
Sourabh Betigeri 850a18bf2f SWDEV-396656 - Adds CG tiled_partition and coalesced_groups tests 
- Tests meta_group_rank/size() in static tile
               partitioning

Change-Id: I1a0d83224145a2356aa12b5e83b094834b53d5a4
 2023-09-22 04:49:19 +00:00
Ioannis Assiouras fb17aa5c41 SWDEV-401489 - [ABI Break] Removed hipModuleLaunchKernelExt 
Change-Id: I0b31d9db755f8a89ead203dbd238a93c18b76855


[ROCm/clr commit: fe2d747bcc]
 2023-09-21 23:50:16 -04:00
Ioannis Assiouras fe2d747bcc SWDEV-401489 - [ABI Break] Removed hipModuleLaunchKernelExt 
Change-Id: I0b31d9db755f8a89ead203dbd238a93c18b76855
 2023-09-21 23:50:16 -04:00
sdashmiz 6f1ac2aa53 SWDEV-330665 - [ABI Break] Add missing enum for attr 
- add hipDevAttrHostRegisterSupported

Signed-off-by: sdashmiz <shadi.dashmiz@amd.com>
Change-Id: I3b176c3463f81ec36b783c81973fa1517abe3a5d


[ROCm/hip commit: f33cb2bb5c]
 2023-09-21 23:42:52 -04:00
sdashmiz f33cb2bb5c SWDEV-330665 - [ABI Break] Add missing enum for attr 
- add hipDevAttrHostRegisterSupported

Signed-off-by: sdashmiz <shadi.dashmiz@amd.com>
Change-Id: I3b176c3463f81ec36b783c81973fa1517abe3a5d
 2023-09-21 23:42:52 -04:00
Jonathan R. Madsen ba4c3a123e Update build for rocprofiler-library to support unit tests (#76) 
* Update build for rocprofiler-library

- Build rocprofiler OBJECT library (rocprofiler-object-library)
- Generate rocprofiler shared library via rocprofiler-object-library
- Generate rocprofiler static library via rocprofiler-object-library
  - this target is excluded from all target and, thus, is only built when another target links to it

* Update lib/rocprofiler/CMakeLists.txt

- tweak order of EXCLUDE_FROM_ALL and STATIC in add_library(rocprofiler-static-library ...)

* Update samples

- link against rocprofiler::rocprofiler instead of rocprofiler::rocprofiler-library following target renaming

* Update cmake

- fix PROJECT_BUILD_TARGETS for rocprofiler::rocprofiler target in rocprofiler-config.cmake
- disable <OS>-<VERSION> in CPack package name
- Add ROCPROFILER_BUILD_CODECOV option

[ROCm/rocprofiler-sdk commit: 5c07deb159]
 2023-09-21 18:01:20 -05:00
Jonathan R. Madsen 5c07deb159 Update build for rocprofiler-library to support unit tests (#76) 
* Update build for rocprofiler-library

- Build rocprofiler OBJECT library (rocprofiler-object-library)
- Generate rocprofiler shared library via rocprofiler-object-library
- Generate rocprofiler static library via rocprofiler-object-library
  - this target is excluded from all target and, thus, is only built when another target links to it

* Update lib/rocprofiler/CMakeLists.txt

- tweak order of EXCLUDE_FROM_ALL and STATIC in add_library(rocprofiler-static-library ...)

* Update samples

- link against rocprofiler::rocprofiler instead of rocprofiler::rocprofiler-library following target renaming

* Update cmake

- fix PROJECT_BUILD_TARGETS for rocprofiler::rocprofiler target in rocprofiler-config.cmake
- disable <OS>-<VERSION> in CPack package name
- Add ROCPROFILER_BUILD_CODECOV option
 2023-09-21 18:01:20 -05:00
Jonathan R. Madsen 218666ebe9 Linting workflow and clang-tidy fixes (#72) 
* Update source/{bin,lib/{common,rocprofiler}}/CMakeLists.txt

- activate clang-tidy

* Update PTL submodule

- clang-tidy fixes

* Update .clang-tidy

- ignore performance-enum-size

* Update CI workflow

- update paths-ignore

* Add linting workflow

- runs clang-tidy

* Update cmake/rocprofiler_build_settings.cmake

- minor modification of flags not recognized by clang-tidy

* Update samples (all of them)

- rocprofiler-samples-build-flags target with -W -Wall -Wextra -Wshadow [-Werror]
- Link samples targets to rocprofiler-samples-build-flags if target exists
- Remove unused variable in main.cpp of api_{buffered,callback}_tracing
- Update samples/pc_sampling
  - single-user-multiple-agents.cpp ends up with unused function find_first_gpu_agent() error
  - change find_first_gpu_agent to return std::optional<rocprofiler_agent_t>
  - change usage after call to find_first_gpu_agent()
  - use find_first_gpu_agent() in single-user-multiple-agents.cpp to determine if there are any GPUs

* Update linting workflow

- fix path to run-ci.py script

* Update linting workflow

- install cmake

* Update common/container/stable_vector.hpp

- fix clang-tidy warning for readability-container-size-empty

[ROCm/rocprofiler-sdk commit: 34505943b2]
 2023-09-21 14:35:20 -05:00
Jonathan R. Madsen 34505943b2 Linting workflow and clang-tidy fixes (#72) 
* Update source/{bin,lib/{common,rocprofiler}}/CMakeLists.txt

- activate clang-tidy

* Update PTL submodule

- clang-tidy fixes

* Update .clang-tidy

- ignore performance-enum-size

* Update CI workflow

- update paths-ignore

* Add linting workflow

- runs clang-tidy

* Update cmake/rocprofiler_build_settings.cmake

- minor modification of flags not recognized by clang-tidy

* Update samples (all of them)

- rocprofiler-samples-build-flags target with -W -Wall -Wextra -Wshadow [-Werror]
- Link samples targets to rocprofiler-samples-build-flags if target exists
- Remove unused variable in main.cpp of api_{buffered,callback}_tracing
- Update samples/pc_sampling
  - single-user-multiple-agents.cpp ends up with unused function find_first_gpu_agent() error
  - change find_first_gpu_agent to return std::optional<rocprofiler_agent_t>
  - change usage after call to find_first_gpu_agent()
  - use find_first_gpu_agent() in single-user-multiple-agents.cpp to determine if there are any GPUs

* Update linting workflow

- fix path to run-ci.py script

* Update linting workflow

- install cmake

* Update common/container/stable_vector.hpp

- fix clang-tidy warning for readability-container-size-empty
 2023-09-21 14:35:20 -05:00
Jatin Chaudhary 668bcece7a SWDEV-411064 - fix device mem size to accurately match memory size from same device 
Change-Id: I9caee84c58312053a7b387c8516b386d53c2197f


[ROCm/hip-tests commit: c340f47ac1]
 2023-09-21 14:03:02 -04:00
Jatin Chaudhary c340f47ac1 SWDEV-411064 - fix device mem size to accurately match memory size from same device 
Change-Id: I9caee84c58312053a7b387c8516b386d53c2197f
 2023-09-21 14:03:02 -04:00
kjayapra-amd 142492d9eb SWDEV-422580 - Adding pcie.function to the PCI address string in hipGetDevicePCIBusId. 
Change-Id: Iaa73b6028ac7b1d05cc310405d6fe622d764320f


[ROCm/clr commit: 369fdc4d87]
 2023-09-21 12:22:05 -04:00
kjayapra-amd 369fdc4d87 SWDEV-422580 - Adding pcie.function to the PCI address string in hipGetDevicePCIBusId. 
Change-Id: Iaa73b6028ac7b1d05cc310405d6fe622d764320f
 2023-09-21 12:22:05 -04:00
Ranjith Ramakrishnan 0dec196bc1 SWDEV-415414 - Corrected the IMPORT_PREFIX path in FindHIP.cmake 
The IMPORT_PREFIX path in FindHIP.cmake was generated as <Install_Prefix>/hip rather than <Install_Prefix> in Linux OS
With file reorg backward compatibility turned off, the path <Install_Prefix>/hip is no more valid
Corrected the same, so that IMPORT_PREFIX will always be generated as <Install_Prefix>

Change-Id: I60b8ee5085c1b4eda0a8494d7f8e1e55accd2f82


[ROCm/hip commit: a3e8fc7a96]
 2023-09-21 12:08:08 -04:00
Ranjith Ramakrishnan a3e8fc7a96 SWDEV-415414 - Corrected the IMPORT_PREFIX path in FindHIP.cmake 
The IMPORT_PREFIX path in FindHIP.cmake was generated as <Install_Prefix>/hip rather than <Install_Prefix> in Linux OS
With file reorg backward compatibility turned off, the path <Install_Prefix>/hip is no more valid
Corrected the same, so that IMPORT_PREFIX will always be generated as <Install_Prefix>

Change-Id: I60b8ee5085c1b4eda0a8494d7f8e1e55accd2f82
 2023-09-21 12:08:08 -04:00
Sourabh Betigeri 5d160eb7ee SWDEV-326798 - [ABI Break] Fixes null stream sync behavior 
Change-Id: I84ad21f61779145c198dc68bf0fe02a57d34bf64


[ROCm/clr commit: 3f88fe850b]
 2023-09-21 11:57:58 -04:00
Sourabh Betigeri 3f88fe850b SWDEV-326798 - [ABI Break] Fixes null stream sync behavior 
Change-Id: I84ad21f61779145c198dc68bf0fe02a57d34bf64
 2023-09-21 11:57:58 -04:00
Jatin Chaudhary acc203fd8d SWDEV-422808 - Add E5M2 and E4M3 for nvidia headers 
Change-Id: Ib40e75c1b2e18d75164607a46e95fc322363c08e


[ROCm/hipother commit: 2eb5e93d26]
 2023-09-21 11:57:04 -04:00
Jatin Chaudhary 2eb5e93d26 SWDEV-422808 - Add E5M2 and E4M3 for nvidia headers 
Change-Id: Ib40e75c1b2e18d75164607a46e95fc322363c08e
 2023-09-21 11:57:04 -04:00
Jatin Chaudhary 6abb81112d SWDEV-422808 - Add E5M2 and E4M3 for nvidia headers 
Change-Id: Ib40e75c1b2e18d75164607a46e95fc322363c08e


[ROCm/clr commit: 15ba5c0004]
 2023-09-21 11:57:04 -04:00
Jatin Chaudhary 15ba5c0004 SWDEV-422808 - Add E5M2 and E4M3 for nvidia headers 
Change-Id: Ib40e75c1b2e18d75164607a46e95fc322363c08e
 2023-09-21 11:57:04 -04:00
Satyanvesh Dittakavi be8dbcf736 SWDEV-375607 - Fix hipGraphExecUpdate behavior 
Addresses the below scenarios like parameters mismatch
in memcopy node, difference in the count of nodes, difference
in the dependencies of the nodes.

Change-Id: I31c6516fb27cc1007809f1b50306fdb0c2356ccc


[ROCm/clr commit: f16d336e32]
 2023-09-21 11:17:49 -04:00
Satyanvesh Dittakavi f16d336e32 SWDEV-375607 - Fix hipGraphExecUpdate behavior 
Addresses the below scenarios like parameters mismatch
in memcopy node, difference in the count of nodes, difference
in the dependencies of the nodes.

Change-Id: I31c6516fb27cc1007809f1b50306fdb0c2356ccc
 2023-09-21 11:17:49 -04:00
Jonathan R. Madsen ac832eb2c7 Fix rocprofiler installation (#73) 
- install rocprofiler library
- define AMD_INTERNAL_BUILD when including hsa/hsa.h
- install include/rocprofiler/registration.h header
- fix samples/pc_sampling cmake via installed rocprofiler
- fix samples/api_callback_tracing cmake via installed rocprofiler
- fix samples/api_buffered_tracing cmake via installed rocprofiler
- set cmake_minimum_required in samples/CMakeLists.txt
- find dependent packages in rocprofiler-config.cmake.in
  - AMDDeviceLibs
  - amd_comgr
  - hsa-runtime64
  - hip
- export rocprofiler-hip and rocprofiler-hsa-runtime libraries
- add Test Install Build step to CI workflow

[ROCm/rocprofiler-sdk commit: 8be4ca1a04]
 2023-09-21 10:07:07 -05:00
Jonathan R. Madsen 8be4ca1a04 Fix rocprofiler installation (#73) 
- install rocprofiler library
- define AMD_INTERNAL_BUILD when including hsa/hsa.h
- install include/rocprofiler/registration.h header
- fix samples/pc_sampling cmake via installed rocprofiler
- fix samples/api_callback_tracing cmake via installed rocprofiler
- fix samples/api_buffered_tracing cmake via installed rocprofiler
- set cmake_minimum_required in samples/CMakeLists.txt
- find dependent packages in rocprofiler-config.cmake.in
  - AMDDeviceLibs
  - amd_comgr
  - hsa-runtime64
  - hip
- export rocprofiler-hip and rocprofiler-hsa-runtime libraries
- add Test Install Build step to CI workflow
 2023-09-21 10:07:07 -05:00
Oliveira, Daniel 1bf68ad1c9 rocm_smi_lib: Fix [linux BM] [AMDSMI] Memory Bandwidth 
Implements APIs for 'gpu_metrics_v1_3' utilization averages

Code changes related to the following:
  * rsmi_dev_activity_metric_get()
  * rsmi_dev_activity_avg_mm_get()
  * CLI shows "Avg.Memory Bandwidth" under "--showmemuse"

Change-Id: I8e4600f350a7c18499abf022534db2b875f09d5f
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>


[ROCm/rocm_smi_lib commit: e0483f2ee2]
 2023-09-21 11:00:29 -04:00
Oliveira, Daniel e0483f2ee2 rocm_smi_lib: Fix [linux BM] [AMDSMI] Memory Bandwidth 
Implements APIs for 'gpu_metrics_v1_3' utilization averages

Code changes related to the following:
  * rsmi_dev_activity_metric_get()
  * rsmi_dev_activity_avg_mm_get()
  * CLI shows "Avg.Memory Bandwidth" under "--showmemuse"

Change-Id: I8e4600f350a7c18499abf022534db2b875f09d5f
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
 2023-09-21 11:00:29 -04:00
Jatin Chaudhary cb193f70c4 SWDEV-411064 - check if image is supported before running image specific tests 
Change-Id: I62b3f7489f2f33b4b4e02f42a46e66c8940f2277


[ROCm/hip-tests commit: ee3c94faf2]
 2023-09-21 06:23:55 -04:00
Jatin Chaudhary ee3c94faf2 SWDEV-411064 - check if image is supported before running image specific tests 
Change-Id: I62b3f7489f2f33b4b4e02f42a46e66c8940f2277
 2023-09-21 06:23:55 -04:00
Rakesh Roy 6369d2a789 SWDEV-2 - Bump HIP version 
- Bump HIP version from 5.7 to 6.0

Change-Id: I95f36d9bb14d470b633394c65238327163cf849b


[ROCm/hip commit: d4c04f46f1]
 2023-09-21 06:11:18 -04:00
Rakesh Roy d4c04f46f1 SWDEV-2 - Bump HIP version 
- Bump HIP version from 5.7 to 6.0

Change-Id: I95f36d9bb14d470b633394c65238327163cf849b
 2023-09-21 06:11:18 -04:00
Satyanvesh Dittakavi 5e0f45df53 SWDEV-419150 - Address incorrect wave size check for gfx7 
Change-Id: Iebb15d3165dc4203fd0c3dda6db785d00f65bbfd


[ROCm/hip commit: 9902b4aed6]
 2023-09-21 03:20:16 -04:00
Satyanvesh Dittakavi 9902b4aed6 SWDEV-419150 - Address incorrect wave size check for gfx7 
Change-Id: Iebb15d3165dc4203fd0c3dda6db785d00f65bbfd
 2023-09-21 03:20:16 -04:00
dependabot[bot] 80f60788f7 Bump actions/configure-pages from 2 to 3 (#68) 
* Bump actions/configure-pages from 2 to 3

Bumps [actions/configure-pages](https://github.com/actions/configure-pages) from 2 to 3.
- [Release notes](https://github.com/actions/configure-pages/releases)
- [Commits](https://github.com/actions/configure-pages/compare/v2...v3)

---
updated-dependencies:
- dependency-name: actions/configure-pages
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>

* Update scripts/thread-sanitizer-suppr.txt

- replace race_top with race since it appears that race_top isn't suppressing the thread sanitizer error from libamdhip64.so

---------

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>

[ROCm/rocprofiler-sdk commit: 78425069e8]
 2023-09-21 02:17:17 -05:00
dependabot[bot] 78425069e8 Bump actions/configure-pages from 2 to 3 (#68) 
* Bump actions/configure-pages from 2 to 3

Bumps [actions/configure-pages](https://github.com/actions/configure-pages) from 2 to 3.
- [Release notes](https://github.com/actions/configure-pages/releases)
- [Commits](https://github.com/actions/configure-pages/compare/v2...v3)

---
updated-dependencies:
- dependency-name: actions/configure-pages
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>

* Update scripts/thread-sanitizer-suppr.txt

- replace race_top with race since it appears that race_top isn't suppressing the thread sanitizer error from libamdhip64.so

---------

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>
 2023-09-21 02:17:17 -05:00