MCPcopy Create free account
hub / github.com/DeepRec-AI/DeepRec / DoBlasGemv

Method DoBlasGemv

tensorflow/stream_executor/cuda/cuda_blas.cc:1085–1093  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

1083}
1084
1085bool CUDABlas::DoBlasGemv(Stream *stream, blas::Transpose trans, uint64 m,
1086 uint64 n, float alpha, const DeviceMemory<float> &a,
1087 int lda, const DeviceMemory<float> &x, int incx,
1088 float beta, DeviceMemory<float> *y, int incy) {
1089 return DoBlasInternal<float>(
1090 cublasSgemv, stream, true /* = pointer_mode_host */,
1091 CUDABlasTranspose(trans), m, n, &alpha, GpuMemory(a), lda, GpuMemory(x),
1092 incx, &beta, GpuMemoryMutable(y), incy);
1093}
1094
1095bool CUDABlas::DoBlasGemv(Stream *stream, blas::Transpose trans, uint64 m,
1096 uint64 n, double alpha, const DeviceMemory<double> &a,

Callers

nothing calls this directly

Calls 4

CUDABlasTransposeFunction · 0.85
GpuMemoryFunction · 0.85
GpuMemoryMutableFunction · 0.85
GpuComplexFunction · 0.85

Tested by

no test coverage detected