* Optimize RDC counter sampling with greedy packing algorithm
This change significantly reduces the number of rocprofiler-sdk sample calls
by implementing a greedy packing algorithm that groups multiple counters into
the minimal number of hardware profiles.
Key improvements:
- Implement greedy packing algorithm to combine counters into minimal profiles
- Add ProfileSet structure to manage packed counter configurations
- Cache packed profile sets for reuse across queries
- Group telemetry field requests by GPU for bulk processing
- Reduce sample calls by ~35% (from 100 to 65 for typical workloads)
Performance impact:
- 13 counters now packed into 3 profiles (77% compression)
- Reduces overhead from profile creation and context switching
- More efficient utilization of hardware counter resources
Implementation details:
- Added create_profiles_for_counters() using greedy algorithm
- Added sample_counters_with_packing() for bulk sampling
- Modified telemetry layer to use rocp_lookup_bulk()
- Preserves all field transformations and special handling
Testing shows successful packing with expected performance gains.
No functional changes to external APIs or behavior.
Co-Authored-By: Ben Welton <bwelton@amd.com>
* Address PR review feedback
This commit addresses all review comments from the initial PR:
1. Fix division by zero risk in debug logging
- Added check for empty counters vector before calculating compression ratio
- Avoids potential division by zero when logging profile creation stats
2. Improve thread safety for statistics tracking
- Changed static uint64_t to std::atomic<uint64_t> for thread-safe counters
- Prevents race conditions in multi-threaded sampling scenarios
3. Remove unused variable
- Removed unused profile_index variable that was incremented but never used
- Cleaned up dead code
4. Clean up code formatting
- Removed extra blank lines for consistency
- Applied formatting fixes across modified files
5. Refactor code duplication between rocp_lookup and rocp_lookup_bulk
- Created apply_field_transformation() helper function
- Eliminates ~70 lines of duplicated switch statement logic
- Centralizes field transformation logic in single location
- Makes future maintenance easier
6. Document non-rocprofiler metrics handling
- Added comments explaining how bulk lookup handles special cases
- Clarifies that non-profiler fields like KFD_ID are handled in transformation
All changes maintain backward compatibility and pass compilation.
Co-Authored-By: Ben Welton <bwelton@amd.com>
---------
Co-authored-by: Ben Welton <bwelton@amd.com>
Co-authored-by: Adam Pryor <61172547+adam360x@users.noreply.github.com>
NOTE: GPU ordering used is not the same as in HSA/HIP.
GPUs are ordered via amdsmi and then GPU_ID fields are compared to map
GPU partitions to each other.
Change-Id: If379214f5281d7d5ee98515b3e5ba7affc2e2197
Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
[ROCm/rdc commit: 85b619b2f0]
The 'value' pointer was being written to a lot and then used for reading
within the same function. This likely caused issues all over RDC when
reading the metrics.
This commit changes it so *value is written to only once.
Change-Id: I83c158c1e46c6ce46ff87d8a2e769f26ffa8c0da
Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
[ROCm/rdc commit: 91be467cad]
SWDEV-475242
For the description of "FP32 Engine Activity" and "FP64 Engine Activity" in dcgm,
It seems that we do not have an equivalent to these pipe-utilizations on our hardware.
In rocprofiler, I think VALU Utilization is the closest to what we want.
Change-Id: Ibce8835ef4757084cdfd73258de6fc1606ca0158
Signed-off-by: Chen Gong <curry.gong@amd.com>
[ROCm/rdc commit: 251fcbe49d]
- Replace non-working fields with working ones
- remove CU_OCCUPANCY completely as it isn't well supported
- Fix rocprofiler initialization with shared_ptr and rdc_module_init
- Replace env var ROCPROFILER_METRICS_PATH with ROCP_METRICS
- ROCPROFILER_METRICS_PATH is only relevant for rocprofv2
- ROCP_METRICS is only relevant for rocprofv1 (which we are using)
Change-Id: I21e6fa3f0e1694c38f44ca0e5659d672559f7380
Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
[ROCm/rdc commit: 20ca2ce574]
This commit adds integration with ROCmTools
Additional changes:
- Fix DEB and RPM installation issue when systemd is not present
- Fix typos in rdc.h
- Wrap negative values in parentheses in rdc.h
- CMAKE: Improve rocm_smi searching
- README: Improve formatting, add info about ROCmTools
Metrics added: 700-714
Metrics can be listed with `rdci dmon --list-all`
Majority of the metrics are only supported by Instict (MI) series GPUs
700 RDC_FI_PROF_ELAPSED_CYCLES should be available on most devices
See README for more information
Change-Id: I907d3eacdc92fc5588ca6c76c2fa1ce0ad900770
Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
[ROCm/rdc commit: 861a843ed7]
Adam Pryor
Benjamin Welton
Dmitrii
Galantsev, Dmitrii
Pryor, Adam
Greg Scaffidi
Chen Gong