AlpinDale e14223dce5 kernel: use `cub::BlockReduce` instead of custom impl (#895) | 1 bulan lalu | |
---|---|---|
.. | ||
amd | f1d0b77c92 [0.6.0] Release Candidate (#481) | 5 bulan lalu |
nvidia | f1d0b77c92 [0.6.0] Release Candidate (#481) | 5 bulan lalu |
common.cu | e14223dce5 kernel: use `cub::BlockReduce` instead of custom impl (#895) | 1 bulan lalu |
fp8_marlin.cu | f1d0b77c92 [0.6.0] Release Candidate (#481) | 5 bulan lalu |