ciflow/xpu/194864: [inductor] Use Inline ASM HOP in MXFP8 nested reduction test
- PyTorch 近 90 天出现 121 次
- PyTorch 近 90 天第 120 次发版
- 上一次：同一天稍早 · trunk/63e23d5995803a63721873b92e3f6389d9bf5731: Fix grammatical errors in comments and docstrings (#194565)
发生了什么
The MXFP8 nested-reduction test directly called the private cvt_e8m0_rceil prim, whose lowering only supports NVIDIA SM100+. This made the test fail on XPU and expanded the proposed fix into an unnecessary cross-device lowering. Use Inline ASM HOP on CUDA and ordinary tensor bit operations on XPU. This keeps the direct prim unchanged while exercising the same fused nested-reduction kernel on both backends. Test Plan…
摘要按规则整理自下方来源原文