geadd_internal.h

Go to the documentation of this file.

#pragma once
#include "cblas.h"
#include "tensor/tensor.h"

namespace magmadnn {
namespace internal {

template <typename T>
bool geadd_check(Tensor<T> *A, Tensor<T> *B, Tensor<T> *C);

template <typename T>
void geadd_full(T alpha, Tensor<T> *A, T beta, Tensor<T> *B, Tensor<T> *C);

#if defined(_HAS_CUDA_)

template <typename T>
void geadd_full_device(T alpha, Tensor<T> *A, T beta, Tensor<T> *B, Tensor<T> *C);
#endif


template <typename T>
void tensor_scalar_add_full(T alpha, Tensor<T> *x, Tensor<T> *out);

#if defined(_HAS_CUDA_)

template <typename T>
void tensor_scalar_add_full_device(T alpha, Tensor<T> *x, Tensor<T> *out);
#endif

}   // namespace internal
}   // namespace magmadnn