AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
8,920 Commits 3 Branches 0 Tags
93958aa8b9a8b0c19ea57ed15cddc668bef0ffa9
Commit Graph

8920 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Jose Santos 597acee21c remove extra saved_dfs test 
Signed-off-by: Jose Santos <josantos@amd.com>
 2024-02-28 14:17:42 -06:00
Jonathan Kim ed462035fa Disable SDMA ganging on non-APU multi-partition modes 
Work around SDMA hang in non-SPX modes for non-APU devices by disabling
ganging.
Root cause of hang not found.
non-APU xGMI modes have only 1 link between socket devices anyways so
there's likely no real system level gain in ganging intra-socket.

Change-Id: Ia4eda2f85cbf25151d3dbcf50cc45b8b775c60e2
 2024-02-28 14:52:01 -05:00
Karl W Schulz 7c6b965851 testing code coverage summary 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-28 11:54:12 -06:00
Jose Santos 8a15b991de reformatting 
Signed-off-by: Jose Santos <josantos@login1.hpcfund>
 2024-02-28 10:31:22 -06:00
Jose Santos 5515cd20e9 Add testing save-dfs in analyze mode 
Signed-off-by: Jose Santos <josantos@login1.hpcfund>
 2024-02-28 10:31:22 -06:00
Karl W Schulz 0fb3796145 fix yaml syntax 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 17:20:07 -06:00
Karl W Schulz 186d5a078a reduce testing threadcount 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 17:16:34 -06:00
Karl W Schulz 9aee84cb43 re-enable updated analyze_commands tests - now configured to run in parallel for 
extra fun

Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 17:08:05 -06:00
Karl W Schulz 9b63bbaff1 bump minimum python version to 3.8 for 2.x 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 15:11:17 -06:00
Karl W Schulz 593d08a4f3 update test matrix to start with ROCm 5.7.1 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 15:11:17 -06:00
Karl W Schulz 2a6285a16b document updated minimum version requirements for ROCm and python 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 15:11:17 -06:00
Karl W Schulz 1adc67bc9e re-enable updated analyze_workload tests - now running in parallel 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 14:05:13 -06:00
Karl W Schulz ec0e7ff22e add PYTEST_NUMPROCS option to define number of CPU threads for pytest 
where appropriate; defaults to a value of 1 but can be overwridden for
CI purposed

Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 13:46:33 -06:00
Giovanni LB 4711a083de SWDEV-445711: Workaround for race condition on exit 
Change-Id: I7391637005ccc7ea89611f28691c02e569197dff
 2024-02-27 14:45:47 -05:00
Jonathan Kim ed260ea970 Fix gang item wait on dependency signals 
Gang items have to wait on dependency signals as well as the leader.
Copies should not start if shaders are still operating on memory
to be copied.

Change-Id: I99703b420045ebcba2c9da39ec64678129dc140f
 2024-02-27 12:45:41 -05:00
Karl W Schulz 85e9462f2a apply formatting pass 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 11:36:00 -06:00
Karl W Schulz 53cf8973d2 re-order to sort functions 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 11:36:00 -06:00
Karl W Schulz 95c6d9d98c update analyze_workload test names for consistency with other pytest files 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 11:36:00 -06:00
Karl W Schulz ec88842ce2 add pytest-xdist to testing requirements 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 11:36:00 -06:00
Karl W Schulz 980524dda5 disable older test publisher 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 10:39:12 -06:00
Karl W Schulz c1c401f03e update path for junit results 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 10:23:36 -06:00
Karl W Schulz fab5f4bd51 enable alternative junit publisher 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 10:16:04 -06:00
Karl W Schulz 635363dcaf update junit output config for profiler tests 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 10:13:46 -06:00
AravindanC 2d7167d068 SWDEV-428381 Remove Hardcoded rocm path 
Change-Id: I39cfbc9add62813076db1992faea4426638859af
 2024-02-27 10:41:54 -05:00
Karl W Schulz f08e2eebca add pytest-results-action 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 09:27:16 -06:00
Karl W Schulz 84e5b89a9c endeavor to add test summary 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 09:05:39 -06:00
Karl W Schulz 26ff6da78b disable METRIC_COMPARE option 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 09:05:39 -06:00
Karl W Schulz 1df0985b0c update soc checks for MI100 in roofline tests 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-27 08:56:39 -06:00
Karl W Schulz 5a1a3a53c9 fix workloads path for profile tests 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-26 15:08:49 -06:00
Karl W Schulz f3ec1f08b4 reorder one metric_threshold entry 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-26 14:48:33 -06:00
Jose Santos d4b910d359 updating Baseline workload directory 
Signed-off-by: Jose Santos <josantos@t006-001.hpcfund>
 2024-02-26 14:48:33 -06:00
Jose Santos ea9ab293e7 update test_profile_general.py 
Signed-off-by: Jose Santos <josantos@login1.hpcfund>
 2024-02-26 14:48:33 -06:00
Jose Santos 23db40d257 Adding pytest_profile_general.py 
Signed-off-by: Jose Santos <josantos@login1.hpcfund>
 2024-02-26 14:48:33 -06:00
Charis Poag a74062cc30 Merge amd-staging into amd-master 20240226 
Change-Id: I1d6db79aa35dabbfb4b837ffdb5dd63ff099cbd9
Signed-off-by: Charis Poag <Charis.Poag@amd.com>
 2024-02-26 13:50:16 -06:00
Harish Kasiviswanathan 341ecaf1d9 libhsakmt: Associate correct GPU with queue memory 
Pass the correct gpu_id to KFD for system memory that is allocated for
the queue and eop buffer

Change-Id: I43bb6333560a7d9d38293c191303161ab1443b5d
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
 2024-02-26 14:20:16 -05:00
Shweta Khatri f2006d6899 Record interop mapped object in allocation_map_ 
This allows the VA to be recorded in ROCr so that they are not
treated as an invalid pointer in future API calls.

Change-Id: I8d1d8ef9816a984c89d30a2179b0ce8940fef1da
 2024-02-26 13:40:55 -05:00
Karl W Schulz f2235b7f1d Fix for rocm-smi not getting mclk for MI100 
Signed-off-by: Nicholas Curtis <nicurtis@amd.com>
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-26 11:47:56 -06:00
Harish Kasiviswanathan 191caf46ac libhsakmt: Move global zfb_support to globals.c 
zfb_support needs to accessed from multiple places, so move to globals.c
file

Change-Id: I40b487c26a13e7cc6fc01b671d6166e7114e02d2
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
 2024-02-26 09:08:11 -05:00
Jonathan R. Madsen e2db37ca8d Search for rocprofiler_set_api_table(...) via RTLD_DEFAULT (#56) 
* Search for rocprofiler_set_api_table(...) via RTLD_DEFAULT

- check if rocprofiler_set_api_table function has already been loaded via dlsym(RTLD_DEFAULT, ...)
- fix warning about conflicting ROCM_DEP_ROCMCORE and ROCPROFILER_REGISTER_DEP_ROCMCORE

* Bump version to 0.4.0
 2024-02-25 21:14:27 -06:00
SrirakshaNag bd81e6a5f8 tools support for callback counter collection (#515) 
* tools support for callback counter collection

* source formatting (clang-format v11) (#516)

Co-authored-by: SrirakshaNag <SrirakshaNag@users.noreply.github.com>

* Fixing conflicts with main

* source formatting (clang-format v11) (#517)

Co-authored-by: SrirakshaNag <SrirakshaNag@users.noreply.github.com>

* source formatting (clang-format v11) (#529)

Co-authored-by: SrirakshaNag <104580803+SrirakshaNag@users.noreply.github.com>

* Adding a column for counter name and counter value

* source formatting (clang-format v11) (#530)

Co-authored-by: SrirakshaNag <104580803+SrirakshaNag@users.noreply.github.com>

* fixing issues with atomic

---------

Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: SrirakshaNag <SrirakshaNag@users.noreply.github.com>
 2024-02-25 20:30:38 -06:00
Harish Kasiviswanathan 858dfd364f libhsakmt: Use correct gpu_id for GPU system memory 
For GTT memory allocation if GPU is provided honour it.

Change-Id: Iea9a26bc44cd3daa2337845f53dc430787b0643b
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
 2024-02-25 10:45:36 -05:00
Ammar ELWazir 4bb95f885b Update run-ci.py (#534) 2024-02-24 14:48:41 -06:00
Jose Santos 5a26c0c87a update test_profile_general.py 
Signed-off-by: Jose Santos <josantos@login1.hpcfund>
 2024-02-23 17:21:38 -06:00
Jose Santos 61c110ccd7 Adding pytest_profile_general.py 
Signed-off-by: Jose Santos <josantos@login1.hpcfund>
 2024-02-23 17:21:38 -06:00
Nicholas Curtis 99a6a244ac Add requirements.txt for docs 
Signed-off-by: Nicholas Curtis <nicurtis@amd.com>
 2024-02-23 15:18:34 -06:00
Jonathan R. Madsen 875f53b608 Correlation ID Retirement + misc (#527) 
* Correlation ID Retirement

- include/rocprofiler-sdk/buffer_tracing.h
  - add rocprofiler_buffer_tracing_correlation_id_retirement_record_t
- include/rocprofiler-sdk/fwd.h
  - ROCPROFILER_BUFFER_TRACING_CORRELATION_ID_RETIREMENT
- lib/rocprofiler-sdk/buffer_tracing.cpp
  - kind string for correlation id retirement
- lib/rocprofiler-sdk/buffer.hpp
  - emplace returns bool
- lib/rocprofiler-sdk/registration.cpp
  - pass lib_instance to copy_table functions
- lib/rocprofiler-sdk/context/context.*
  - update correlation_id struct
    - make ref_count private
    - {get,add,sub}_ref_count() functions
      - sub_ref_count() performs correlation id retirement
    - use stack for "latest" thread-local correlation id
- lib/rocprofiler-sdk/hip/hip.*
  - migrate to new {get,add,sub}_ref_count() for correlation ids
  - return in iterate_args
  - handle table instance in copy_table
- lib/rocprofiler-sdk/hsa/hsa.*
  - migrate to new {get,add,sub}_ref_count() for correlation ids
  - return in iterate_args
  - handle table instance in copy_table
- lib/rocprofiler-sdk/marker/marker.*
  - migrate to new {get,add,sub}_ref_count() for correlation ids
  - return in iterate_args
  - handle table instance in copy_table
- lib/rocprofiler-sdk/hsa/async_copy.cpp
  - migrate to new {get,add,sub}_ref_count() for correlation ids
  - handle table instance in async_copy_init / async_copy_save
- lib/rocprofiler-sdk/hsa/queue.cpp
  - migrate to new {get,add,sub}_ref_count() for correlation ids
  - tweak to external correlation id mapping in WriteInterceptor
- tests/async-copy-tracing/validate.py
  - check retired_correlation_ids
- tests/common/serialization.hpp
  - support rocprofiler_buffer_tracing_correlation_id_retirement_record_t
- tests/kernel-tracing/validate.py
  - check retired_correlation_ids
- tests/common/CMakeLists.txt
  - perfetto external project
- tests/common/perfetto.hpp
  - perfetto categories + aliases
  - add_perfetto_annotation
  - metaprogramming helpers
- tests/tools/CMakeLists.txt
  - link to tests-perfetto
- tests/tools/json-tool.cpp
  - demangling functions
  - serialization of marker API callback args
  - reduce parallel bottleneck in tool_tracing_callback
  - support correlation id retirement
  - Multiple threads for buffers
  - Support ROCPROFILER_TOOL_CONTEXTS_EXCLUDE env variable
  - write_perfetto() function

* Update tests/rocprofv3/tracing/validate.py

- tweak test_hsa_api_trace

* Update PTL submodule

- fixes for data race during destruction of task

* Update lib/rocprofiler-sdk/buffer.*

- unique_buffer_vec_t uses std::unique_ptr instead of allocator::unique_static_ptr_t

* Reduce timeouts in counter collection samples [skip ci]

* Update tests/tools/json-tool.cpp

- tweak demangle(string_view, int*) -> demangle(string_view, int&)

* Update lib/rocprofiler-sdk/hsa/async_copy.cpp

- move sub_ref_count() to later in async_copy_handler to delay retirement slightly more
 2024-02-23 10:30:33 -06:00
Jonathan R. Madsen 1f0d06b6cb Correct the install locations of cmake config files (#55) 
* Correct the install locations of cmake config files

The cmake config files should be installed in <Install_Prefix>/lib/cmake/<component_name>

* Apply suggestions from code review

* Correct the install locations of cmake config files

- ROCm standard for cmake config files is <install-prefix>/lib/cmake/<package-name>

---------

Co-authored-by: Ranjith Ramakrishnan <Ranjith.Ramakrishnan@amd.com>
 2024-02-23 10:17:58 -06:00
Karl W Schulz aeb2161b2a updating testing files to meet formatting requirements 
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>
 2024-02-23 10:00:23 -06:00
Jose Santos b018d9207f updating Pytest mark and tests/workloads 
Signed-off-by: Jose Santos <josantos@login1.hpcfund>
 2024-02-23 10:00:23 -06:00
Jose Santos 4f96837157 updating analyze commands to new commands/workloads 
Signed-off-by: Jose Santos <josantos@login1.hpcfund>
 2024-02-23 10:00:23 -06:00