GitLab

Summary

The gpu_stream benchmark has NVIDIA-specific dependencies that prevent
it from compiling on ROCm 6.3+. This change makes it CUDA-only,
gracefully skipping the build with a warning on non-NVIDIA
  environments.

  Problem

The gpu_stream benchmark fails to compile on ROCm 6.3+ due to multiple
NVIDIA-specific dependencies:

1. nvml.h — NVIDIA Management Library header, used for querying actual
memory clock rates. No HIP equivalent. Referenced in gpu_stream.cu and
gpu_stream_utils.hpp.
2. cuda.h in headers — Three .hpp files (gpu_stream.hpp,
gpu_stream_kernels.hpp, gpu_stream_utils.hpp) directly include <cuda.h>
and <cuda_runtime.h>. These headers are not processed by hipify-perl
(only
  .cu source files are), so they fail to resolve on ROCm.
3. Deprecated hipDeviceProp_t struct fields — The code accesses
memoryBusWidth, memoryClockRate, and ECCEnabled from the device
properties struct. These fields were removed from hipDeviceProp_t in
ROCm
    6.3, causing compilation errors after hipification.

The existing ROCm path was marked as incomplete (# TODO: test for ROC)
and was never fully functional on recent ROCm versions.

  Changes

- Removed the non-functional ROCm/HIP build path from
gpu_stream/CMakeLists.txt
- When CUDA is not found, prints a warning and returns gracefully
instead of attempting a broken hipify build or raising FATAL_ERROR
- No changes to the NVIDIA/CUDA build path — it continues to work as
before

  Impact

   - NVIDIA builds: No change — gpu_stream builds and installs normally
- ROCm builds: gpu_stream is skipped with a warning message. Previously
it would fail the entire make cppbuild step, blocking the Docker image
build
- Other benchmarks: Unaffected — build.sh continues to the next
benchmark after gpu_stream returns
Co-authored-by: Hongtao Zhang <hongtaozhang@microsoft.com>

Added GPU Stream benchmark - measures the GPU memory bandwidth and
efficiency for double datatype through various memory operations
including copy, scale, add, and triad.
- added documentation for `gpu-stream` detailing its introduction,
metrics, and descriptions.
- added unit tests for `gpu-stream`. Example output is in
`superbenchmark/tests/data/gpu_stream.log`.

Benchmarks: Micro benchmark - Add one-to-all, all-to-one, all-to-all support to gpu_copy_bw_performance (#588)

**Description**
Add one-to-all, all-to-one, all-to-all support to
gpu_copy_bw_performance, and fix performance bug in gpu_copy

**Description**
Use config `log_raw_data` to control whether log the raw data into file or not. The default value is `no`. We can set it as `yes` for some particular benchmarks to save the raw data into file, such as NCCL/RCCL test.

This commit makes data checking in gpu_copy optional, because it will take too long time if message size is large.

**Description**
This commit does the following to optimize result variance in gpu_copy benchmark:
1) Add warmup phase for gpu_copy benchmark to avoid timing instability caused by first-time CUDA kernel launch overhead;
2) Use CUDA events for timing instead of CPU timestamps;
3) Make data checking an option that is not preferred to be enabled in performance test;
4) Enlarge message size in performance benchmark.

**Description**
This commit adds bidirectional tests in gpu_copy benchmark for both device-host transfer and device-device transfer, and revises related tests.

**Description**
Unify metric names of benchmarks.

**Description**
This commit does the following:
1) Adds CPU-initiated copy benchmark;
2) Adds dtod benchmark;
3) Support scanning NUMA nodes and GPUs inside the benchmark program;
4) Change the name of gpu-sm-copy to gpu-copy.

**Description**
This commit adds gpu_sm_copy benchmark and related tests.

**Description**
Rename `kernel_launch_overhead_event` to `event_overhead`, `kernel_launch_overhead_wall` to `wall_overhead`.

* add kernel launch overhead benchmark.