[ Mixed ] fix apply using casted function

[platform/core/ml/nntrainer.git] / nntrainer / tensor / blas_interface.h

diff --git a/nntrainer/tensor/blas_interface.h b/nntrainer/tensor/blas_interface.h
index b09d945..0193e4d 100644 (file)
--- a/nntrainer/tensor/blas_interface.h
+++ b/nntrainer/tensor/blas_interface.h
@@ -2,12 +2,12 @@
 /**
  * Copyright (C) 2020 Jijoong Moon <jijoong.moon@samsung.com>
  *
- * @file       blas_interface.h
- * @date       28 Aug 2020
- * @see                https://github.com/nnstreamer/nntrainer
- * @author     Jijoong Moon <jijoong.moon@samsung.com>
- * @bug                No known bugs except for NYI items
- * @brief      This is dummy header for blas support
+ * @file   blas_interface.h
+ * @date   28 Aug 2020
+ * @see    https://github.com/nnstreamer/nntrainer
+ * @author Jijoong Moon <jijoong.moon@samsung.com>
+ * @bug    No known bugs except for NYI items
+ * @brief  This is dummy header for blas support
  *
  */
 
@@ -27,6 +27,7 @@ enum CBLAS_TRANSPOSE {
   CblasTrans = 112,
   CblasConjTrans = 113
 };
+
 #endif
 
 #ifdef USE_CUBLAS
@@ -34,32 +35,77 @@ enum CBLAS_TRANSPOSE {
 #include <helper_functions.h>
 #endif
 
+#include <tensor_dim.h>
+
 namespace nntrainer {
 
-void sscal(const int N, const float alpha, float *X, const int incX);
+#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);
 
 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);
 
 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);
 
+unsigned int isamax(const unsigned int N, const float *X, const int incX);
+
 } /* namespace nntrainer */
 #endif /* __cplusplus */
 #endif /* __BLAS_INTERFACE_H__ */