namespace nntrainer {
+#ifdef ENABLE_FP16
+void sscal(const unsigned int N, const float alpha, _FP16 *X, const int incX);
+_FP16 snrm2(const int N, const _FP16 *X, const int incX);
+void scopy(const unsigned int N, const _FP16 *X, const int incX, _FP16 *Y,
+ const int intY);
+_FP16 sdot(const unsigned int N, const _FP16 *X, const unsigned int incX,
+ const _FP16 *Y, const unsigned int incY);
+void saxpy(const unsigned int N, const float alpha, const _FP16 *X,
+ const int incX, _FP16 *Y, const int incY);
+void sgemm(CBLAS_ORDER order, CBLAS_TRANSPOSE TransA, CBLAS_TRANSPOSE TransB,
+ const unsigned int M, const unsigned int N, const unsigned int K,
+ const float alpha, const _FP16 *A, const unsigned int lda,
+ const _FP16 *B, const unsigned int ldb, const float beta, _FP16 *C,
+ const unsigned int ldc);
+void sgemv(CBLAS_ORDER order, CBLAS_TRANSPOSE TransA, const unsigned int M,
+ const unsigned int N, const float alpha, const _FP16 *A,
+ const unsigned int lda, const _FP16 *X, const int incX,
+ const float beta, _FP16 *Y, const int incY);
+unsigned int isamax(const unsigned int N, const _FP16 *X, const int incX);
+#endif
+
void sscal(const unsigned int N, const float alpha, void *X, const int incX,
ml::train::TensorDim::DataType d_type);
void sscal(const unsigned int N, const float alpha, float *X, const int incX);
-void sscal(const unsigned int N, const float alpha, __fp16 *X, const int incX);
-
float snrm2(const int N, const float *X, const int incX);
void scopy(const unsigned int N, const void *X, const int incX, void *Y,
const int incY, ml::train::TensorDim::DataType d_type);
-
void scopy(const unsigned int N, const float *X, const int incX, float *Y,
const int intY);
-void scopy(const unsigned int N, const __fp16 *X, const int incX, __fp16 *Y,
- const int intY);
-
float sdot(const unsigned int N, const float *X, const unsigned int incX,
const float *Y, const unsigned int incY);
+void saxpy(const unsigned int N, const float alpha, const void *X,
+ const int incX, void *Y, const int incY,
+ ml::train::TensorDim::DataType d_type);
void saxpy(const unsigned int N, const float alpha, const float *X,
const int incX, float *Y, const int incY);
void sgemm(CBLAS_ORDER order, CBLAS_TRANSPOSE TransA, CBLAS_TRANSPOSE TransB,
const unsigned int M, const unsigned int N, const unsigned int K,
+ const float alpha, const void *A, const unsigned int lda,
+ const void *B, const unsigned int ldb, const float beta, void *C,
+ const unsigned int ldc, ml::train::TensorDim::DataType d_type);
+
+void sgemm(CBLAS_ORDER order, CBLAS_TRANSPOSE TransA, CBLAS_TRANSPOSE TransB,
+ const unsigned int M, const unsigned int N, const unsigned int K,
const float alpha, const float *A, const unsigned int lda,
const float *B, const unsigned int ldb, const float beta, float *C,
const unsigned int ldc);
void sgemv(CBLAS_ORDER order, CBLAS_TRANSPOSE TransA, const unsigned int M,
+ const unsigned int N, const float alpha, const void *A,
+ const unsigned int lda, const void *X, const int incX,
+ const float beta, void *Y, const int incY,
+ ml::train::TensorDim::DataType d_type);
+
+void sgemv(CBLAS_ORDER order, CBLAS_TRANSPOSE TransA, const unsigned int M,
const unsigned int N, const float alpha, const float *A,
const unsigned int lda, const float *X, const int incX,
const float beta, float *Y, const int incY);