LLM inference in C/C++
C++ ggml
metal : fix moe_reduce fusion race when dst aliases its inputs (#30795)
Georgi Gerganov098ce526b546
diff --git a/ggml/src/ggml-metal/ggml-metal-fusion.cpp b/ggml/src/ggml-metal/ggml-metal-fusion.cppindex d78ca7c4b7..020cad29d1 100644--- a/ggml/src/ggml-metal/ggml-metal-fusion.cpp+++ b/ggml/src/ggml-metal/ggml-metal-fusion.cpp@@ -636,6 +636,11 @@ static bool ggml_metal_fusion_check_moe_reduce(if (!match.experts->data || !match.weights->data || !match.dst->data) {return false;}++ // the fused kernel reads experts and weights while writing dst, so dst must not overlap either input+ if (ggml_metal_fusion_overlap(match.dst, match.experts) || ggml_metal_fusion_overlap(match.dst, match.weights)) {+ return false;+ }}return true;