tail-hip / README.md
pat883's picture
Add kernel card (library_name: kernels)
c60ff2c verified
|
Raw
History Blame Contribute Delete
577 Bytes
metadata
library_name: kernels
tags:
  - kernels
  - rocm

Fused Elementwise Tail Ops (RDNA4/gfx1201)

Fused RMSNorm (+residual), SiLU-and-mul, and RoPE native HIP kernels.

Built with kernel-builder for AMD RDNA4 (gfx1201). Load with the kernels library:

from kernels import get_kernel
kernel = get_kernel("pat883/tail-hip")

Requires a ROCm PyTorch build (torch 2.10 / ROCm 7.x) on an RDNA4 card. Built variants: torch210-cxx11-rocm70 and torch210-cxx11-rocm71.