trunk/4144bea4b7c2f67da9d4840f4659f977bbd754e5: [MPS] Remove svd_kernel_mps small-input CPU gate (#195950)
- PyTorch 近 90 天出现 602 次
- PyTorch 近 90 天第 596 次发版
- 上一次：同一天稍早 · trunk/13948faea12f1eb3a6df72f181a51f3e65dbf309: [MPS] Orthonormalize null-space U columns in svd_jacobi (#196139)
发生了什么
svd_kernel_mps routed inputs with numel infos.any().item(), a mandatory device->host sync on every call regardless of backend, so the GPU path does not pipeline and there is no async barrier for the gate to avoid. The CPU fallback then adds two data transfers (A to host, U/S/Vh back) on top of that same sync, making it strictly slower. Measured on M5 Pro, native GPU path vs the exact work the gate did (A.to(cpu) ->…
摘要按规则整理自下方来源原文