Post
PyTorch
pytorch.org
did:plc:o3pmmtxoor73pszydv7pffsv
Helion, a PyTorch-native kernel DSL, powers a vLLM linear backend with per-shape autotuning. On Hopper, tuning plus hybrid dispatch beat vLLM’s default CUTLASS and DeepGEMM backends.
https://pytorch.org/blog/building-a-high-performance-and-portable-vllm-linear-backend-with-helion/
2026-10-02T20:05:21.866Z