Skip to content

Kernel Optimization Guide ​

This guide explains the optimization techniques used in TensorCraft Core kernels.

GEMM Optimization Journey ​

Key Techniques ​

  • Shared Memory Tiling
  • Double Buffering
  • Register Blocking
  • Tensor Core (WMMA)
  • MMA PTX
  • Software Pipelining

References ​

Released under the MIT License.