Open Issues Need Help
View All on GitHub b200 / gb200 database entries, once the numbers can be cross-checked about 1 hour ago
help wanted
Query every CUDA device attribute without profiling, and benchmark kernels against your hardware's theoretical peak.
Python
#benchmarking#cuda#gpu#nsight-compute#triton
windows: confirm nvcuda.dll and nvml.dll load paths about 1 hour ago
help wanted
Query every CUDA device attribute without profiling, and benchmark kernels against your hardware's theoretical peak.
Python
#benchmarking#cuda#gpu#nsight-compute#triton
compare: accept custom cards from a local json about 2 hours ago
enhancement good first issue
Query every CUDA device attribute without profiling, and benchmark kernels against your hardware's theoretical peak.
Python
#benchmarking#cuda#gpu#nsight-compute#triton
add the rtx 4070 family to the gpu database about 2 hours ago
good first issue
Query every CUDA device attribute without profiling, and benchmark kernels against your hardware's theoretical peak.
Python
#benchmarking#cuda#gpu#nsight-compute#triton