Gopesh Bhardwaj
348d740388
* hsa multiqueuw application * cmake formatting (cmake-format) (#619) Co-authored-by: bgopesh <7112102+bgopesh@users.noreply.github.com> * source formatting (clang-format v11) (#620) Co-authored-by: bgopesh <7112102+bgopesh@users.noreply.github.com> * comppialtion fix * Update tests/bin/CMakeLists.txt Reorder `add_subdirectory` to fix (recurrent) issues with ROCTx in `CMAKE_BUILD_RPATH` * addressing early feedback * cmake updates * more cmake updates * adding queue dependency test * updating test * test updates * removed hsa_api_trace header * reformating headers to prevent clang from reordering * Fixing packaging * Fixes for hangs * source formatting (clang-format v11) (#676) Co-authored-by: bwelton <1683479+bwelton@users.noreply.github.com> * cmake formatting (cmake-format) (#673) Co-authored-by: bwelton <1683479+bwelton@users.noreply.github.com> * structure change * cmake formatting (cmake-format) (#680) Co-authored-by: bgopesh <7112102+bgopesh@users.noreply.github.com> * Adding kernel trace to test hang fix * rebased and fixed kernel-trace test * rhel clang fixes * source formatting (clang-format v11) (#685) Co-authored-by: bgopesh <7112102+bgopesh@users.noreply.github.com> * Update lib/rocprofiler-sdk-tool/helper.hpp - remove suppression of -Wshadow * Update tests/bin/hsa-queue-dependency/CMakeLists.txt - cleanup unnecessary code - GPU_LIST -> GPU_TARGETS * GPU_LIST -> GPU_TARGETS * Remove installation of test executable and libraries --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: bgopesh <7112102+bgopesh@users.noreply.github.com> Co-authored-by: Jonathan R. Madsen <jrmadsen@users.noreply.github.com> Co-authored-by: Benjamin Welton <bewelton@amd.com> Co-authored-by: bwelton <1683479+bwelton@users.noreply.github.com> Co-authored-by: Jonathan R. Madsen <jonathanrmadsen@gmail.com>
68 řádky
2.0 KiB
Markdown
68 řádky
2.0 KiB
Markdown
# ROCprofiler-SDK: Application Profiling, Tracing, and Performance Analysis
|
||
|
||
***
|
||
Note: rocprofiler-sdk is currently `not` supported as part of public ROCm software stack and only distributed as a beta
|
||
release to customers.
|
||
***
|
||
|
||
## Overview
|
||
|
||
ROCProfiler-SDK is AMD’s new and improved tooling infrastructure that provides a hardware specific low level performance analysis interface for the profiling and the tracing of GPU compute applications. To see whats changed [Click Here](source/docs/about.md)
|
||
|
||
## GPU Metrics
|
||
|
||
- GPU hardware counters
|
||
- HIP API tracing
|
||
- HIP kernel tracing
|
||
- HSA API tracing
|
||
- HSA operation tracing
|
||
- Marker(ROCtx) tracing
|
||
|
||
## Tool Support
|
||
|
||
rocprofv3 is the command line tool that is built using rocprofiler-sdk library and shipped with ROCm stack. To see details on
|
||
the command line options of rocprofv3, please see rocprofv3 user guide
|
||
[Click Here](source/docs/rocprofv3.md)
|
||
|
||
## Documentation
|
||
|
||
We make use of doxygen to automatically generate API documentation. Generated document can be found in the following path:
|
||
|
||
``` bash
|
||
<ROCM_PATH>/share/html/rocprofiler-sdk
|
||
```
|
||
|
||
ROCM_PATH by default is /opt/rocm
|
||
It can be set by the user in different location if needed.
|
||
|
||
## Build and Installation
|
||
|
||
```bash
|
||
git clone https://git@github.com:ROCm/rocprofiler-sdk-internal.git rocprofiler-sdk-source
|
||
cmake \
|
||
-B rocprofiler-sdk-build \
|
||
-D ROCPROFILER_BUILD_TESTS=ON \
|
||
-D ROCPROFILER_BUILD_SAMPLES=ON \
|
||
-D ROCPROFILER_BUILD_DOCS=ON \
|
||
-D CMAKE_INSTALL_PREFIX=/opt/rocm \
|
||
rocprofiler-sdk-source
|
||
|
||
cmake --build rocprofiler-sdk-build --target all --parallel 8
|
||
```
|
||
|
||
To install ROCprofiler, run:
|
||
|
||
```bash
|
||
cmake --build rocprofiler-sdk-build --target install
|
||
```
|
||
|
||
Please see detailed section on build and installation here: [Click Here](/source/docs/installation.md)
|
||
|
||
## Support
|
||
|
||
Please report in the Github Issues.
|
||
|
||
## Limitations
|
||
|
||
- Individual xcc mode is not supported.
|