gemm_sse.h File Reference

Generalized matrix matrix multiplication using SSE intrinsics. More...

#include <stdexcept>
#include "mm_kernel_inner_sse2_A.h"
#include "mm_kernel_outer_A.h"

Go to the source code of this file.

Functions
template<typename real, typename regType, int m_kernel, int n_kernel, int k_kernel, int m_block, int n_block>
static void	gemm_sse (real const *const A, real const *const B, real *C, size_t const m, size_t const n, size_t const k, real *A_packed, real *B_packed, real *C_packed, size_t const ap_size, size_t const bp_size, size_t const cp_size)
template<typename real>
static void	gemm_sse (real const *const A, real const *const B, real *C, size_t const m, size_t const n, size_t const k, real *A_packed, real *B_packed, real *C_packed, size_t const ap_size, size_t const bp_size, size_t const cp_size)
template<>
void	gemm_sse (double const *const A, double const *const B, double *C, size_t const m, size_t const n, size_t const k, double *A_packed, double *B_packed, double *C_packed, size_t const ap_size, size_t const bp_size, size_t const cp_size)
template<>
void	gemm_sse (float const *const A, float const *const B, float *C, size_t const m, size_t const n, size_t const k, float *A_packed, float *B_packed, float *C_packed, size_t const ap_size, size_t const bp_size, size_t const cp_size)

Detailed Description

Generalized matrix matrix multiplication using SSE intrinsics.

Author

Emanuel H. Rubensson

Date

2009

Function Documentation

gemm_sse() [1/4]

template<>

void gemm_sse	(	double const *const	A,
		double const *const	B,
		double *	C,
		size_t const	m,
		size_t const	n,
		size_t const	k,
		double *	A_packed,
		double *	B_packed,
		double *	C_packed,
		size_t const	ap_size,
		size_t const	bp_size,
		size_t const	cp_size )

References A, B, and gemm_sse().

gemm_sse() [2/4]

template<>

void gemm_sse	(	float const *const	A,
		float const *const	B,
		float *	C,
		size_t const	m,
		size_t const	n,
		size_t const	k,
		float *	A_packed,
		float *	B_packed,
		float *	C_packed,
		size_t const	ap_size,
		size_t const	bp_size,
		size_t const	cp_size )

References A, B, and gemm_sse().

gemm_sse() [3/4]

template<typename real, typename regType, int m_kernel, int n_kernel, int k_kernel, int m_block, int n_block>

void gemm_sse ( real const *const A, real const *const B, real * C, size_t const m, size_t const n, size_t const k, real * A_packed, real * B_packed, real * C_packed, size_t const ap_size, size_t const bp_size, size_t const cp_size )

static

References A, and B.

Referenced by mat::gemm(), gemm_sse(), and gemm_sse().

gemm_sse() [4/4]

template<typename real>

void gemm_sse ( real const *const A, real const *const B, real * C, size_t const m, size_t const n, size_t const k, real * A_packed, real * B_packed, real * C_packed, size_t const ap_size, size_t const bp_size, size_t const cp_size )

static

References A, and B.