2x4.c File Reference

Detailed Description

The $2\times 4$ MBCSR implementation of $A^TA\cdot x$ and $A^HA\cdot x$ .

Automatically generated by ./gen_ata.sh on Wed Jun 8 15:55:01 PDT 2005.

#include <assert.h>
#include <oski/config.h>
#include <oski/common.h>
#include <oski/mangle.h>
#include <oski/vecview.h>
#include <oski/MBCSR/format.h>
#include <oski/MBCSR/module.h>

Defines

#define REGISTER   register

Real-valued, so use explicit 'register' keyword.

#define MBCSR_MatTransMatMult_v1_aX_b1_xs1_ys1   MANGLE_MOD_(MBCSR_MatTransMatMult_v1_aX_b1_xs1_ys1_2x4)

Mangled name for MBCSR_MatTransMatMult_v1_aX_b1_xs1_ys1.

#define MBCSR_MatTransMatMult_v1_aX_b1_xsX_ysX   MANGLE_MOD_(MBCSR_MatTransMatMult_v1_aX_b1_xsX_ysX_2x4)

Mangled name for MBCSR_MatTransMatMult_v1_aX_b1_xsX_ysX.

#define SubmatReprTransSubmatReprMult   MANGLE_MOD_(SubmatReprTransSubmatReprMult_2x4)

Mangled name for primary exportable symbol, SubmatReprTransSubmatReprMult.

#define MBCSR_MatHermMatMult_v1_aX_b1_xs1_ys1   MANGLE_MOD_(MBCSR_MatHermMatMult_v1_aX_b1_xs1_ys1_2x4)

Mangled name for MBCSR_MatHermMatMult_v1_aX_b1_xs1_ys1.

#define MBCSR_MatHermMatMult_v1_aX_b1_xsX_ysX   MANGLE_MOD_(MBCSR_MatHermMatMult_v1_aX_b1_xsX_ysX_2x4)

Mangled name for MBCSR_MatHermMatMult_v1_aX_b1_xsX_ysX.

#define SubmatReprHermSubmatReprMult   MANGLE_MOD_(SubmatReprHermSubmatReprMult_2x4)

Mangled name for primary exportable symbol, SubmatReprHermSubmatReprMult.

Functions

void MBCSR_MatTransMatMult_v1_aX_b1_xs1_ys1 (oski_index_t M, const oski_index_t *restrict ptr, const oski_index_t *restrict ind, const oski_value_t *restrict val, const oski_value_t *restrict diag, oski_value_t alpha, const oski_value_t *restrict x, oski_value_t *restrict y, oski_value_t *restrict t, oski_index_t inct)

The $2\times 4$ MBCSR implementation of $A^TA\cdot x$ , where the source and destination vectors have unit-stride and unit-stride, respectively.

void MBCSR_MatTransMatMult_v1_aX_b1_xsX_ysX (oski_index_t M, const oski_index_t *restrict ptr, const oski_index_t *restrict ind, const oski_value_t *restrict val, const oski_value_t *restrict diag, oski_value_t alpha, const oski_value_t *restrict x, oski_index_t incx, oski_value_t *restrict y, oski_index_t incy, oski_value_t *restrict t, oski_index_t inct)

The $2\times 4$ MBCSR implementation of $A^TA\cdot x$ , where the source and destination vectors have general-stride and general-stride, respectively.

int SubmatReprTransSubmatReprMult (const oski_submatMBCSR_t *A, oski_value_t alpha, const oski_vecview_t x, oski_vecview_t y, oski_vecview_t t)

Exported module wrapper for the $2\times 4$ implementation of $A^TA\cdot x$ .

void MBCSR_MatHermMatMult_v1_aX_b1_xs1_ys1 (oski_index_t M, const oski_index_t *restrict ptr, const oski_index_t *restrict ind, const oski_value_t *restrict val, const oski_value_t *restrict diag, oski_value_t alpha, const oski_value_t *restrict x, oski_value_t *restrict y, oski_value_t *restrict t, oski_index_t inct)

The $2\times 4$ MBCSR implementation of $A^HA\cdot x$ , where the source and destination vectors have unit-stride and unit-stride, respectively.

void MBCSR_MatHermMatMult_v1_aX_b1_xsX_ysX (oski_index_t M, const oski_index_t *restrict ptr, const oski_index_t *restrict ind, const oski_value_t *restrict val, const oski_value_t *restrict diag, oski_value_t alpha, const oski_value_t *restrict x, oski_index_t incx, oski_value_t *restrict y, oski_index_t incy, oski_value_t *restrict t, oski_index_t inct)

The $2\times 4$ MBCSR implementation of $A^HA\cdot x$ , where the source and destination vectors have general-stride and general-stride, respectively.

int SubmatReprHermSubmatReprMult (const oski_submatMBCSR_t *A, oski_value_t alpha, const oski_vecview_t x, oski_vecview_t y, oski_vecview_t t)

Exported module wrapper for the $2\times 4$ implementation of $A^HA\cdot x$ .

Function Documentation

void MBCSR_MatHermMatMult_v1_aX_b1_xs1_ys1 ( oski_index_t M,

const oski_index_t *restrict ptr,

const oski_index_t *restrict ind,

const oski_value_t *restrict val,

const oski_value_t *restrict diag,

oski_value_t alpha,

const oski_value_t *restrict x,

oski_value_t *restrict y,

oski_value_t *restrict t,

oski_index_t inct

)

The $2\times 4$ MBCSR implementation of $A^HA\cdot x$ , where the source and destination vectors have unit-stride and unit-stride, respectively.
Parameter 'tp' may be NULL if no intermediate vector output is desired.

void MBCSR_MatHermMatMult_v1_aX_b1_xsX_ysX ( oski_index_t M,

const oski_index_t *restrict ptr,

const oski_index_t *restrict ind,

const oski_value_t *restrict val,

const oski_value_t *restrict diag,

oski_value_t alpha,

const oski_value_t *restrict x,

oski_index_t incx,

oski_value_t *restrict y,

oski_index_t incy,

oski_value_t *restrict t,

oski_index_t inct

)

The $2\times 4$ MBCSR implementation of $A^HA\cdot x$ , where the source and destination vectors have general-stride and general-stride, respectively.
Parameter 'tp' may be NULL if no intermediate vector output is desired.

void MBCSR_MatTransMatMult_v1_aX_b1_xs1_ys1 ( oski_index_t M,

const oski_index_t *restrict ptr,

const oski_index_t *restrict ind,

const oski_value_t *restrict val,

const oski_value_t *restrict diag,

oski_value_t alpha,

const oski_value_t *restrict x,

oski_value_t *restrict y,

oski_value_t *restrict t,

oski_index_t inct

)

The $2\times 4$ MBCSR implementation of $A^TA\cdot x$ , where the source and destination vectors have unit-stride and unit-stride, respectively.
Parameter 'tp' may be NULL if no intermediate vector output is desired.

void MBCSR_MatTransMatMult_v1_aX_b1_xsX_ysX ( oski_index_t M,

const oski_index_t *restrict ptr,

const oski_index_t *restrict ind,

const oski_value_t *restrict val,

const oski_value_t *restrict diag,

oski_value_t alpha,

const oski_value_t *restrict x,

oski_index_t incx,

oski_value_t *restrict y,

oski_index_t incy,

oski_value_t *restrict t,

oski_index_t inct

)

The $2\times 4$ MBCSR implementation of $A^TA\cdot x$ , where the source and destination vectors have general-stride and general-stride, respectively.
Parameter 'tp' may be NULL if no intermediate vector output is desired.

Generated on Wed Sep 19 16:41:19 2007 for BeBOP Optimized Sparse Kernel Interface Library by

1.4.6


Defines
#define	REGISTER register
	Real-valued, so use explicit 'register' keyword.
#define	MBCSR_MatTransMatMult_v1_aX_b1_xs1_ys1 MANGLE_MOD_(MBCSR_MatTransMatMult_v1_aX_b1_xs1_ys1_2x4)
	Mangled name for MBCSR_MatTransMatMult_v1_aX_b1_xs1_ys1.
#define	MBCSR_MatTransMatMult_v1_aX_b1_xsX_ysX MANGLE_MOD_(MBCSR_MatTransMatMult_v1_aX_b1_xsX_ysX_2x4)
	Mangled name for MBCSR_MatTransMatMult_v1_aX_b1_xsX_ysX.
#define	SubmatReprTransSubmatReprMult MANGLE_MOD_(SubmatReprTransSubmatReprMult_2x4)
	Mangled name for primary exportable symbol, SubmatReprTransSubmatReprMult.
#define	MBCSR_MatHermMatMult_v1_aX_b1_xs1_ys1 MANGLE_MOD_(MBCSR_MatHermMatMult_v1_aX_b1_xs1_ys1_2x4)
	Mangled name for MBCSR_MatHermMatMult_v1_aX_b1_xs1_ys1.
#define	MBCSR_MatHermMatMult_v1_aX_b1_xsX_ysX MANGLE_MOD_(MBCSR_MatHermMatMult_v1_aX_b1_xsX_ysX_2x4)
	Mangled name for MBCSR_MatHermMatMult_v1_aX_b1_xsX_ysX.
#define	SubmatReprHermSubmatReprMult MANGLE_MOD_(SubmatReprHermSubmatReprMult_2x4)
	Mangled name for primary exportable symbol, SubmatReprHermSubmatReprMult.
Functions
void	MBCSR_MatTransMatMult_v1_aX_b1_xs1_ys1 (oski_index_t M, const oski_index_t restrict ptr, const oski_index_t restrict ind, const oski_value_t restrict val, const oski_value_t restrict diag, oski_value_t alpha, const oski_value_t restrict x, oski_value_t restrict y, oski_value_t *restrict t, oski_index_t inct)
	The $2\times 4$ MBCSR implementation of $A^TA\cdot x$ , where the source and destination vectors have unit-stride and unit-stride, respectively.
void	MBCSR_MatTransMatMult_v1_aX_b1_xsX_ysX (oski_index_t M, const oski_index_t restrict ptr, const oski_index_t restrict ind, const oski_value_t restrict val, const oski_value_t restrict diag, oski_value_t alpha, const oski_value_t restrict x, oski_index_t incx, oski_value_t restrict y, oski_index_t incy, oski_value_t *restrict t, oski_index_t inct)
	The $2\times 4$ MBCSR implementation of $A^TA\cdot x$ , where the source and destination vectors have general-stride and general-stride, respectively.
int	SubmatReprTransSubmatReprMult (const oski_submatMBCSR_t *A, oski_value_t alpha, const oski_vecview_t x, oski_vecview_t y, oski_vecview_t t)
	Exported module wrapper for the $2\times 4$ implementation of $A^TA\cdot x$ .
void	MBCSR_MatHermMatMult_v1_aX_b1_xs1_ys1 (oski_index_t M, const oski_index_t restrict ptr, const oski_index_t restrict ind, const oski_value_t restrict val, const oski_value_t restrict diag, oski_value_t alpha, const oski_value_t restrict x, oski_value_t restrict y, oski_value_t *restrict t, oski_index_t inct)
	The $2\times 4$ MBCSR implementation of $A^HA\cdot x$ , where the source and destination vectors have unit-stride and unit-stride, respectively.
void	MBCSR_MatHermMatMult_v1_aX_b1_xsX_ysX (oski_index_t M, const oski_index_t restrict ptr, const oski_index_t restrict ind, const oski_value_t restrict val, const oski_value_t restrict diag, oski_value_t alpha, const oski_value_t restrict x, oski_index_t incx, oski_value_t restrict y, oski_index_t incy, oski_value_t *restrict t, oski_index_t inct)
	The $2\times 4$ MBCSR implementation of $A^HA\cdot x$ , where the source and destination vectors have general-stride and general-stride, respectively.
int	SubmatReprHermSubmatReprMult (const oski_submatMBCSR_t *A, oski_value_t alpha, const oski_vecview_t x, oski_vecview_t y, oski_vecview_t t)
	Exported module wrapper for the $2\times 4$ implementation of $A^HA\cdot x$ .