Skip to content

LLM inference in C/C++

C++ ggml

CUDA: add cublasSgemmBatched mapping for HIP/MUSA vendor headers (#25033)

leonardHONG2f18fe13c5dd
diff --git a/ggml/src/ggml-cuda/vendors/hip.h b/ggml/src/ggml-cuda/vendors/hip.h
index a6115cd80d..d01f1533ab 100644
--- a/ggml/src/ggml-cuda/vendors/hip.h
+++ b/ggml/src/ggml-cuda/vendors/hip.h
@@ -48,6 +48,7 @@
#define cublasSetMathMode(handle, mode) CUBLAS_STATUS_SUCCESS
#define cublasSetStream hipblasSetStream
#define cublasSgemm hipblasSgemm
+#define cublasSgemmBatched hipblasSgemmBatched
#define cublasSgemmStridedBatched hipblasSgemmStridedBatched
#define cublasStatus_t hipblasStatus_t
#define cublasOperation_t hipblasOperation_t
diff --git a/ggml/src/ggml-cuda/vendors/musa.h b/ggml/src/ggml-cuda/vendors/musa.h
index 99e8fa3703..6d725c7ec1 100644
--- a/ggml/src/ggml-cuda/vendors/musa.h
+++ b/ggml/src/ggml-cuda/vendors/musa.h
@@ -32,6 +32,7 @@
#define cublasSetMathMode mublasSetMathMode
#define cublasSetStream mublasSetStream
#define cublasSgemm mublasSgemm
+#define cublasSgemmBatched mublasSgemmBatched
#define cublasSgemmStridedBatched mublasSgemmStridedBatched
#define cublasStatus_t mublasStatus_t
#define cublasOperation_t mublasOperation_t