nvbench
nvbench copied to clipboard
NVIDIA
CUDA Kernel Benchmarking Library
To implement a Setup, which is shared among all states of a `benchmark_base` a current solution could be: ```cpp void my_benchmark(nvbench::state& state) { static int num_execs = 0; if (execs...
To implement Global Setup and TearDown feature, similar to [GoogleTest Global Set-Up and Tear-Down](https://github.com/google/googletest/blob/main/docs/advanced.md#global-set-up-and-tear-down). There are use cases in [cudf ](https://github.com/rapidsai/cudf/blob/acb6aed5bc5929d7ec97a30a3fbfebd86bf0d703/cpp/benchmarks/io/orc/orc_reader.cpp#L199-L203)that could use Global Set-Up and Tear-Down feature, as pointed...
Add a `--color` option to `scripts/nvbench_compare.py`, and only import/use colorama when this is explicitly passed in.
Adding NVML as an optional dependency would allow some cool features: * Lock clock frequency. * Per-device default frequency. * Per-device maximum frequency. * Explicit frequency. * Log various device...
Right now it looks like benchmark names must be the full name of the benchmark, but it would be useful to provide a regex syntax as gbench allows.
We should add an `output` directory under `nvbench/example` that contains the json, csv, and md outputs for each benchmark on a multidevice system. This should be linked to from the...
NVBench has a work-in-progress JSON output format and I'm working on a very basic python script to compare two JSON files. We should grow this functionality into a more complete...
# Overview RAPIDS needs a way to iterate through correlated axes in lockstep. # Example Usecase Consider a benchmark that takes three parameters "X", "Y", and "Z", where X is...
There are some compilation errors with CUPTI from CTK 11.2: ``` FAILED: nvbench/CMakeFiles/nvbench.dir/cupti_profiler.cxx.o /home/av/code/utils/ccache-install/bin/g++ -DFMT_LOCALE -Dnvbench_EXPORTS -I/home/av/code/src/nvbench -I/home/av/code/build/nvbench-ctk_11.2 -I/home/av/code/build/nvbench-ctk_11.2/_deps/fmt-src/include -I/home/av/code/build/nvbench-ctk_11.2/nvbench/detail -isystem /usr/local/cuda_11.2.0_460.22/include -isystem /usr/local/cuda_11.2.0_460.22/extras/CUPTI/include -isystem /home/av/code/build/nvbench-ctk_11.2/_deps/nlohmann_json-src/include -O3 -DNDEBUG -fPIC -Wall...
Adds samples of `--list`, `--md`, `--json`, and `--csv` for each example. Fixed a race condition where calling `std::exit(0)` in the parser was skipping ofstream destructors. Some files were never flushed/closed,...