single mention · inference & serving
Split-K CuTe GEMM
A low-latency split-K CuTe GEMM used for small-M Hyper-Connection Mix.
The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See inference kernel for the methods filed beside it.
Also called low-latency split-K CuTe GEMM.
Filed under inference & serving › inference kernel