OpenBLAS/kernel/arm/cgemm_ncopy_2_vfp.S

259 lines
5.3 KiB
ArmAsm

/***************************************************************************
Copyright (c) 2013, The OpenBLAS Project
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are
met:
1. Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in
the documentation and/or other materials provided with the
distribution.
3. Neither the name of the OpenBLAS project nor the names of
its contributors may be used to endorse or promote products
derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE OPENBLAS PROJECT OR CONTRIBUTORS BE
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*****************************************************************************/
/**************************************************************************************
* 2013/11/05 Saar
* BLASTEST : OK
* CTEST : OK
* TEST : OK
*
**************************************************************************************/
#define ASSEMBLER
#include "common.h"
#define STACKSIZE 256
#define OLD_M r0
#define OLD_N r1
#define OLD_A r2
#define OLD_LDA r3
/******************************************************
* [fp, #-128] - [fp, #-64] is reserved
* for store and restore of floating point
* registers
*******************************************************/
#define LDA [fp, #-260 ]
#define B [fp, #4 ]
#define M r0
#define N r1
#define A r2
#define BO r5
#define AO1 r6
#define AO2 r7
#define I r3
#define J r12
#define A_PRE 256
/**************************************************************************************
* Macro definitions
**************************************************************************************/
.macro COPY2x2
flds s0 , [ AO1, #0 ]
flds s1 , [ AO1, #4 ]
flds s4 , [ AO1, #8 ]
flds s5 , [ AO1, #12 ]
flds s2 , [ AO2, #0 ]
flds s3 , [ AO2, #4 ]
add AO1, AO1, #16
flds s6 , [ AO2, #8 ]
flds s7 , [ AO2, #12 ]
vstmia.f32 BO!, { s0 - s7 }
add AO2, AO2, #16
.endm
.macro COPY1x2
flds s0 , [ AO1, #0 ]
flds s1 , [ AO1, #4 ]
flds s2 , [ AO2, #0 ]
flds s3 , [ AO2, #4 ]
add AO1, AO1, #8
vstmia.f32 BO!, { s0 - s3 }
add AO2, AO2, #8
.endm
.macro COPY2x1
flds s0 , [ AO1, #0 ]
flds s1 , [ AO1, #4 ]
flds s2 , [ AO1, #8 ]
flds s3 , [ AO1, #12 ]
vstmia.f32 BO!, { s0 - s3 }
add AO1, AO1, #16
.endm
.macro COPY1x1
flds s0 , [ AO1, #0 ]
flds s1 , [ AO1, #4 ]
vstmia.f32 BO!, { s0 - s1 }
add AO1, AO1, #8
.endm
/**************************************************************************************
* End of macro definitions
**************************************************************************************/
PROLOGUE
.align 5
push {r4 - r9, fp}
add fp, sp, #24
sub sp, sp, #STACKSIZE // reserve stack
lsl r3, r3, #3 // lda = lda * 4 * 2
str r3, LDA
sub r4, fp, #128
vstm r4, { s8 - s15} // store floating point registers
ldr BO, B
/*********************************************************************************************/
cgemm_ncopy_L2_BEGIN:
asrs J, N, #1 // J = N / 2
ble cgemm_ncopy_L1_BEGIN
cgemm_ncopy_L2_M2_BEGIN:
mov AO1, A // AO1 = A
ldr r4 , LDA
add AO2, AO1, r4
add A , AO2, r4 // A = A + 2 * LDA
asrs I, M, #1 // I = M / 2
ble cgemm_ncopy_L2_M2_40
cgemm_ncopy_L2_M2_20:
pld [ AO1, #A_PRE ]
pld [ AO2, #A_PRE ]
COPY2x2
subs I , I , #1
ble cgemm_ncopy_L2_M2_40
COPY2x2
subs I , I , #1
bne cgemm_ncopy_L2_M2_20
cgemm_ncopy_L2_M2_40:
ands I, M , #1
ble cgemm_ncopy_L2_M2_END
cgemm_ncopy_L2_M2_60:
COPY1x2
subs I , I , #1
bne cgemm_ncopy_L2_M2_60
cgemm_ncopy_L2_M2_END:
subs J , J, #1 // j--
bne cgemm_ncopy_L2_M2_BEGIN
/*********************************************************************************************/
cgemm_ncopy_L1_BEGIN:
tst N, #1
ble cgemm_ncopy_L999
cgemm_ncopy_L1_M2_BEGIN:
mov AO1, A // AO1 = A
ldr r4 , LDA
add A , AO1, r4 // A = A + 1 * LDA
asrs I, M, #1 // I = M / 2
ble cgemm_ncopy_L1_M2_40
cgemm_ncopy_L1_M2_20:
COPY2x1
subs I , I , #1
bne cgemm_ncopy_L1_M2_20
cgemm_ncopy_L1_M2_40:
ands I, M , #1
ble cgemm_ncopy_L1_M2_END
cgemm_ncopy_L1_M2_60:
COPY1x1
subs I , I , #1
bne cgemm_ncopy_L1_M2_60
cgemm_ncopy_L1_M2_END:
cgemm_ncopy_L999:
sub r3, fp, #128
vldm r3, { s8 - s15} // restore floating point registers
movs r0, #0 // set return value
sub sp, fp, #24
pop {r4 - r9, fp}
bx lr
EPILOGUE