viable/strict/1791906918: [AARCH64] Use oneDNN jit:int8 matmul for torch._int_mm (#200389)
- PyTorch 近 90 天出现 2262 次
- PyTorch 近 90 天第 2232 次发版
- 上一次：同一天稍早 · viable/strict/1791905525
发生了什么
On AArch64, torch._int_mm always uses the native fallback loop. TorchAO's Int8DynamicActivationInt8WeightConfig calls _int_mm on CPU, so int8 linear layers don't benefit from oneDNN. Since v3.13, oneDNN's AArch64 jit:int8 matmul (SVE + I8MM) supports s32 output ( uxlfoundation/oneDNN#5267 ). This PR routes _int_mm to it. Changes (AArch64 only; x86 unchanged) Dispatch _int_mm to oneDNN when the CPU has SVE and I8MM.…
摘要按规则整理自下方来源原文