* Add OMPT to ROCpd
* Use correct category
* Added wrapper functions for future control
* Formatting
* Fix naming
* Comment change
* Remove ompt_get_cb_args
* Switched to using region_sample for OMPT
* Remove relic function
* Remove get_use_rocpd that was used in this pr (one still remains)
* Rename ompt_get_args_string and reuse in tool_tracing_callback_stop
* Make lock init and destroy cb instant
* [Prototype] ROCPD Name fix
* [Prototype] ROCPD Name fix P1
* [Prototype] ROCPD Name fix P2
* ROCPD Name fix
* Var name changes
* Rewrite cb overwrite to single function
* [Important] Use parallel_data as key for parallel callback map
* Fix workflow failure
* Make cpp USE_ROCM consistent with hpp and use default constructor if USE_ROCM = 0
* Add missing ROCPROFILER_VERSION check
* Improve readability
* Make ompt storage maps thread local
* Part 1: Variable name fix, memory cleanup, and fixed asserts
* Part 2: Add comments
* Part 3: Add CI_THROW
* Part 4: Formatting
* Part 5: Move #include to cpp
Some sections were being displayed multiple times in the web GUI.
Code to append the section was nested inside the subsection loop,
so each time a new subsction was appened to the section,
the entire section was appended.
Added `GPU LINK PORT STATUS` table to `amd-smi xgmi` command
The `amd-smi xgmi -s` or `amd-smi xgmi --source-status` will show `GPU LINK PORT STATUS` table.
---------
Signed-off-by: Bindhiya Kanangot Balakrishnan <Bindhiya.KanangotBalakrishnan@amd.com>
Added `GPU LINK PORT STATUS` table to `amd-smi xgmi` command
The `amd-smi xgmi -s` or `amd-smi xgmi --source-status` will show `GPU LINK PORT STATUS` table.
---------
Signed-off-by: Bindhiya Kanangot Balakrishnan <Bindhiya.KanangotBalakrishnan@amd.com>
[ROCm/amdsmi commit: 7ddd91653e]
Added the following API's to amdsmi_interface.py.
amdsmi_get_cpu_handle()
amdsmi_get_esmi_err_msg()
amdsmi_get_gpu_event_notification()
amdsmi_get_processor_count_from_handles()
amdsmi_get_processor_handles_by_type()
amdsmi_gpu_validate_ras_eeprom()
amdsmi_init_gpu_event_notification()
amdsmi_set_gpu_event_notification_mask()
amdsmi_stop_gpu_event_notification()
amdsmi_get_gpu_busy_percent()
Added additional return value to API amdsmi_get_xgmi_plpd().
The entry policies is added to the end of the dictionary to match API definition.
The entry plpds is marked for deprecation as it has the same information as policies.
---------
Signed-off-by: josnarlo <Joseph.Narlo@amd.com>
Signed-off-by: Maisam Arif <Maisam.Arif@amd.com>
Added the following API's to amdsmi_interface.py.
amdsmi_get_cpu_handle()
amdsmi_get_esmi_err_msg()
amdsmi_get_gpu_event_notification()
amdsmi_get_processor_count_from_handles()
amdsmi_get_processor_handles_by_type()
amdsmi_gpu_validate_ras_eeprom()
amdsmi_init_gpu_event_notification()
amdsmi_set_gpu_event_notification_mask()
amdsmi_stop_gpu_event_notification()
amdsmi_get_gpu_busy_percent()
Added additional return value to API amdsmi_get_xgmi_plpd().
The entry policies is added to the end of the dictionary to match API definition.
The entry plpds is marked for deprecation as it has the same information as policies.
---------
Signed-off-by: josnarlo <Joseph.Narlo@amd.com>
Signed-off-by: Maisam Arif <Maisam.Arif@amd.com>
[ROCm/amdsmi commit: 7decbc67a1]
* Initial steps added for rocprofiler-systems-continuous-integration.yml
* Add new line to end of rocprofiler-systems-continuous-integration.yml
* Fix matrix issue in rocprofiler-systems CI workflow
* Update runner to use mi355
* Remove sudo from ROCm download step
* Add Python venv
* Try to install python venv
* Add -y to pip venv install commands
* Add shell: bash to download ROCm step
* Fix issue in if statement
* Fix typo in mv command
* Fix mv command
* Update paths
* add directory in install step
* Use default runner for now while debugging setup
* Add set -e to steps
* debug build step
* Add amdgpu install step
* remove working-directory from amdgpu install step
* add path/ld lib path, add -S argument to run-ci.py
* Fix typo in DCMAKE_PREFIX_PATH
* Add DGPU_TARGETS to run-ci.py command
* add Docker options, remove GPU_TARGETS
* Install amd-smi-lib
* Add DCMAKE_BUILD_TYPE, update path
* Remove mkdir
* Add build dynist cmake arguments
* Update cmake arguments again
* Add missing \ to run-ci.py command
* add libdw dependency
* Add later install step
* Increase timeout of configure/build/test step
* use 16 jobs to try and speed up pipeline time
* Add GHCR image, remove TheRock tarball download step, minor changes for debugging
* Add credentials to container portion of step
* Add package read permissions to ubuntu step
* Update tarball name
* Increase jobs to 16, disable some tests for now due to timeouts
* Modify to only include gpu tests
* Fix configuration
* Enable MPI on run-ci.py run
* Add install MPI step, changed tests to be run
* Enable OMPI flags, enable network counter access
* Use new Docker image names, add privileged option to Docker
* Change cmake build type
* Add fail-fast false option for CI
* Update ROCM_VERSION variable to reflect docker changes
* Specify TARBALL_ROCM_VERSION as separate
* Add MI325 to debug pipeline errors
* Move location of env variables
* Only test on jammy for now, run all tests to assess other issues
* test with branch that contains fix for openmp
* Exclude "ompvv"
We will re-add one ticket is fixed.
* Test: Disable USE_MPI
* Replace TheRock ROCm install with rocm-dev for now
* Try out MI355 noble and MI325 for jammy/noble
* Update amdgpu step to support different ROCm versions
* Remove unused env variables
---------
Co-authored-by: David Galiffi <David.Galiffi@amd.com>
1) Test linking of SPIRV bundled and unbundled code object in runtime
2) Negative Param test for hipLinkAddData API
Change-Id: I7c8167f6d862f5b23f9532f69b6da4fc50e96dcd
Co-authored-by: Rahul Manocha <rmanocha@amd.com>
* SWDEV-546350 - [catch2][dtest] Test cases for hipMemPrefetchAsync_v2 and hipMemAdvice_v2
* SWDEV-546350 - Enable tests for NVIDIA, Guard failing tests
* SWDEV-546350 - Add checks for Managed Memory Support
* SWDEV-546350 - Gaurded Numa related things against windows
* SWDEV-546350 - Removed some scenarios which needs fixes, gaurded for NVIDIA
* SWDEV-546350 - Incorporated review comments
* SWDEV-546350 - Corrected File name
* SWDEV-546350 - Resetted to default device, modifed initial data filling ways
---------
Co-authored-by: jainprad <92369414+jainprad@users.noreply.github.com>
* SWDEV-546351 - Added test cases for hipGetDriverEntryPoint api
* SWDEV-546351 - Added test cases for hipGetDriverEntryPoint api
* SWDEV-546351 - Added test cases for hipGetDriverEntryPoint api
* SWDEV-546351 : Added test cases for hipGetDriverEntryPoint api
* addressed review comments Added non-zero check
---------
Co-authored-by: Anavena Venkatesh <Anavena.Venkatesh@amd.com>
Co-authored-by: jainprad <92369414+jainprad@users.noreply.github.com>
- Add `nlohmann-json-dev` (or equivalent) to CI Docker images for RHEL, SUSE, and Ubuntu.
- Add `gmock-dev` and `gtest-dev` (or equivalent) to CI Docker images for RHEL, SUSE, and Ubuntu.
- Add `--set solver classic` to conda config to resolve an issue setting up the conda environment
- Fix Perfetto package installation on ubuntu noble image.
- Add a check and log error if pip installation fail
---------
Co-authored-by: jbonnell-amd <jason.bonnell@amd.com>
* rocr: Don't assert in hsa_shut_down when no agents
Instead, print error message and return an error. Prior to
this patch, the assertion would occur when hsa_shut_down() is
called more than once.
* rocr: Reorder Unload ASAN clean-up on shut down
* Used KFD to determine linking between GPUs and PIDs rather than depend on fdinfo's per pid single gpu bdf info that we were getting.
Signed-off-by: adapryor <Adam.pryor@amd.com>
---------
Signed-off-by: adapryor <Adam.pryor@amd.com>
Signed-off-by: Arif, Maisam <Maisam.Arif@amd.com>
* Used KFD to determine linking between GPUs and PIDs rather than depend on fdinfo's per pid single gpu bdf info that we were getting.
Signed-off-by: adapryor <Adam.pryor@amd.com>
---------
Signed-off-by: adapryor <Adam.pryor@amd.com>
Signed-off-by: Arif, Maisam <Maisam.Arif@amd.com>
[ROCm/amdsmi commit: c967aead58]
* Added ability to format gpu_metrics v1_9
* New gpu_metrics format from the driver should allow amd-smi to parse with future compatibility guaranteed
---------
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
Signed-off-by: adapryor <Adam.pryor@amd.com>
Co-authored-by: Oliveira, Daniel <daniel.oliveira@amd.com>
* Added ability to format gpu_metrics v1_9
* New gpu_metrics format from the driver should allow amd-smi to parse with future compatibility guaranteed
---------
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
Signed-off-by: adapryor <Adam.pryor@amd.com>
Co-authored-by: Oliveira, Daniel <daniel.oliveira@amd.com>
[ROCm/amdsmi commit: 5ef0b3c34d]
Pryor, Adam
Oosman Saeed
yalmusaf_amdeng
Kian Cossettini
David Galiffi
German Andryeyev
ammallya
Rahul Manocha
Maisam Arif
Ben Richard
Sam Ruscica
Gopesh Bhardwaj
Kanangot Balakrishnan, Bindhiya
Narlo, Joseph
Venkateshwar Reddy Kandula
Jason Bonnell
ggottipa-amd
systems-assistant[bot]
dependabot[bot]
swargamrambabu
Jaydeep
ajanicijamd
Milan Radosavljevic
vedithal-amd
cfreeamd