2024-11-07 11:51:02 -05:00
# Changelog for ROCm Systems Profiler
Full documentation for ROCm Systems Profiler is available at [https://rocm.docs.amd.com/projects/rocprofiler-systems/en/latest/ ](https://rocm.docs.amd.com/projects/rocprofiler-systems/en/latest/ ).
2025-06-24 11:10:26 -04:00
## ROCm Systems Profiler 1.1.0 for ROCm 7.0
2025-03-06 18:03:33 -05:00
### Added
2025-06-24 11:10:26 -04:00
- Profiling and metric collection capabilities for VCN engine activity, JPEG engine activity, and API tracing for rocDecode, rocJPEG, and VA-APIs.
2025-06-02 13:31:18 -04:00
- How-to document for VCN and JPEG activity sampling and tracing.
2025-06-24 11:10:26 -04:00
- Support for tracing Fortran applications.
- Support for tracing MPI API in Fortran.
2025-07-28 17:33:52 +02:00
- Initial support for rocPD database output with the `ROCPROFSYS_USE_ROCPD` configuration setting.
2025-07-23 21:28:26 -04:00
- By default, group "kernel dispatch" and "memory copy" events by HIP stream ID in Perfetto traces.
- Add the "ROCPROFSYS_ROCM_GROUP_BY_QUEUE" configuration setting to group events by queue, instead.
2025-03-06 18:03:33 -05:00
2025-03-31 11:07:50 -04:00
### Changed
2025-04-24 16:19:13 -04:00
- Replaced ROCm SMI backend with AMD SMI backend for collecting GPU metrics.
2025-06-24 11:10:26 -04:00
- ROCprofiler-SDK is now used to trace RCCL API and collect communication counters.
- Updated the Dyninst submodule to v13.0.
- Set the default value of `ROCPROFSYS_SAMPLING_CPUS` to `none` .
2025-03-31 11:07:50 -04:00
2025-03-20 12:33:48 -04:00
### Resolved issues
2025-06-24 11:10:26 -04:00
- Fixed GPU metric collection settings with `ROCPROFSYS_AMD_SMI_METRICS` .
- Fixed a build issue with CMake 4.
- Fixed incorrect kernel names shown for kernel dispatch tracks in Perfetto.
- Fixed formatting of some output logs.
2025-07-14 19:31:14 -04:00
- Fixed an issue where ROC-TX ranges were displayed as two separate events instead of a single spanning event.
2025-06-24 11:10:26 -04:00
## ROCm Systems Profiler 1.0.2 for ROCm 6.4.2
### Optimized
- Improved readability of the OpenMP target offload traces by showing on a single Perfetto track.
### Resolved issues
- Fixed the file path to the script that merges Perfetto files from multi-process MPI runs. The script has also been renamed from `merge-multiprocess-output.sh` to `rocprof-sys-merge-output.sh` .
2025-03-20 12:33:48 -04:00
2025-05-05 15:06:37 -04:00
## ROCm Systems Profiler 1.0.1 for ROCm 6.4.1
2025-04-24 16:19:13 -04:00
### Added
2025-05-05 15:06:37 -04:00
- How-to document for [network performance profiling ](https://rocm.docs.amd.com/projects/rocprofiler-systems/en/amd-staging/how-to/nic-profiling.html ) for standard Network Interface Cards (NICs).
### Resolved issues
- Fixed a build issue with Dyninst on GCC 13.
2025-04-24 16:19:13 -04:00
## ROCm Systems Profiler 1.0.0 for ROCm 6.4.0
2025-03-20 12:33:48 -04:00
### Added
- Support for VA-API and rocDecode tracing.
2025-04-24 16:19:13 -04:00
- Aggregation of MPI data collected across distributed nodes and ranks. The data is concatenated into a single proto file.
2025-03-20 12:33:48 -04:00
### Changed
2025-04-24 16:19:13 -04:00
- Backend refactored to use ROCprofiler-SDK rather than ROCProfiler and ROCTracer.
2025-03-20 12:33:48 -04:00
### Resolved issues
- Fixed hardware counter summary files not being generated after profiling.
2025-04-24 16:19:13 -04:00
- Fixed an application crash when collecting performance counters with ROCProfiler.
2025-03-20 12:33:48 -04:00
- Fixed interruption in config file generation.
2025-04-24 16:19:13 -04:00
- Fixed segmentation fault while running `rocprof-sys-instrument` .
- Fixed an issue where running `rocprof-sys-causal` or using the `-I all` option with `rocprof-sys-sample` caused the system to become non-responsive.
- Fixed an issue where sampling multi-GPU Python workloads caused the system to stop responding.
2025-03-20 12:33:48 -04:00
2025-01-10 11:05:25 -05:00
## ROCm Systems Profiler 0.1.1 for ROCm 6.3.2
### Resolved issues
- Fixed an error when building from source on some SUSE and RHEL systems when using the `ROCPROFSYS_BUILD_DYNINST` option.
## ROCm Systems Profiler 0.1.0 for ROCm 6.3.1
### Added
- Improvements to support OMPT target offload.
### Resolved issues
- Fixed an issue with generated Perfetto files.
- Fixed an issue with merging multiple `.proto` files.
- Fixed an issue causing GPU resource data to be missing from traces of Instinct MI300A systems.
- Fixed a minor issue for users upgrading to ROCm 6.3 from 6.2 post-rename from `omnitrace` .
2024-11-07 11:51:02 -05:00
## ROCm Systems Profiler 0.1.0 for ROCm 6.3.0
### Changed
- Renamed Omnitrace to ROCm Systems Profiler.
## Omnitrace 1.11.2 for ROCm 6.2.1
### Known issues
- Perfetto can no longer open Omnitrace proto files. Loading the Perfetto trace output `.proto` file in `ui.perfetto.dev` can
result in a dialog with the message, "Oops, something went wrong! Please file a bug." The information in the dialog will
refer to an "Unknown field type." The workaround is to open the files with the previous version of the Perfetto UI found
at https://ui.perfetto.dev/v46.0-35b3d9845/#!/.