OSchip
/
OpenBLAS
Not watched
1
0
Fork 0
Code
Releases 66 Wiki evaluate Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain HPC
Browse Source

LoongArch64: Fixed rot_lsx.S ane crot_lsx.S 

Do not check whether the input parameters c and s are zero,
as this may cause errors with special values (same as scal).
Although OpenBLAS's own test suite doesn't catch this, it will
cause LAPACK test cases to fail.
tags/v0.3.30
gxw 1 year ago
parent
c2212d0abd
commit
e8c740368c
2 changed files with 119 additions and 1883 deletions
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. +6
    -591
      kernel/loongarch64/crot_lsx.S
  2. +113
    -1292
      kernel/loongarch64/rot_lsx.S

+ 6
- 591
kernel/loongarch64/crot_lsx.S View File

@@ -75,6 +75,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
slli.d TEMP, TEMP, ZBASE_SHIFT
slli.d INCX, INCX, ZBASE_SHIFT
slli.d INCY, INCY, ZBASE_SHIFT
move YY, Y
move XX, X
MTG t1, C
MTG t2, S
MTG t3, a1
@@ -89,25 +91,11 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
vreplgr2vr.w VXZ, t3
srai.d I, N, 2
#endif
bge $r0, I, .L997
beq INCX, $r0, .L996
beq INCY, $r0, .L996
bne INCX, TEMP, .L22 // INCX!=1 or INCY!=1
bne INCY, TEMP, .L22

.L11:
bge $r0, I, .L997
CMPEQ $fcc0, C, a1
bcnez $fcc0, .L110
CMPEQ $fcc0, S, a1
bcnez $fcc0, .L112 // C!=0 S==0
b .L111 // C!=0 S!=0
.align 3

.L110:
CMPEQ $fcc0, S, a1
bcnez $fcc0, .L114 // C==0 S==0
b .L113 // C==0 S!=0
.align 3
bne INCX, TEMP, .L221 // INCX!=1 or INCY!=1
bne INCY, TEMP, .L221

.L111: // C!=0 S!=0
vld VX0, X, 0 * SIZE
@@ -168,151 +156,6 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
b .L997
.align 3

.L112: // C!=0 S==0
vld VX0, X, 0 * SIZE
vld VX2, Y, 0 * SIZE
#ifdef DOUBLE
vld VX1, X, 2 * SIZE
vld VX3, Y, 2 * SIZE
vpickev.d x1, VX1, VX0
vpickod.d x2, VX1, VX0
vpickev.d x3, VX3, VX2
vpickod.d x4, VX3, VX2
vfmul.d VX0, x1, VXC
vfmul.d VX1, x3, VXC
vfmul.d VX2, x2, VXC
vfmul.d VX3, x4, VXC
vilvl.d x1, VX2 ,VX0
vilvh.d x2, VX2, VX0
vilvl.d x3, VX3 ,VX1
vilvh.d x4, VX3, VX1
vst x1, X, 0 * SIZE
vst x3, Y, 0 * SIZE
vst x2, X, 2 * SIZE
vst x4, Y, 2 * SIZE
addi.d X, X, 4 * SIZE
addi.d Y, Y, 4 * SIZE
#else
vld VX1, X, 4 * SIZE
vld VX3, Y, 4 * SIZE
vpickev.w x1, VX1, VX0
vpickod.w x2, VX1, VX0
vpickev.w x3, VX3, VX2
vpickod.w x4, VX3, VX2
vfmul.s VX0, x1, VXC
vfmul.s VX1, x3, VXC
vfmul.s VX2, x2, VXC
vfmul.s VX3, x4, VXC
vilvl.w x1, VX2 ,VX0
vilvh.w x2, VX2, VX0
vilvl.w x3, VX3 ,VX1
vilvh.w x4, VX3, VX1
vst x1, X, 0 * SIZE
vst x3, Y, 0 * SIZE
vst x2, X, 4 * SIZE
vst x4, Y, 4 * SIZE
addi.d X, X, 8 * SIZE
addi.d Y, Y, 8 * SIZE
#endif
addi.d I, I, -1
blt $r0, I, .L112
b .L997
.align 3

.L113: // C==0 S!=0
vld VX0, X, 0 * SIZE
vld VX2, Y, 0 * SIZE
#ifdef DOUBLE
vld VX1, X, 2 * SIZE
vld VX3, Y, 2 * SIZE
vpickev.d x1, VX1, VX0
vpickod.d x2, VX1, VX0
vpickev.d x3, VX3, VX2
vpickod.d x4, VX3, VX2
vfmul.d VX0, x3, VXS
vfmul.d VX1, x1, VXS
vfsub.d VX1, VXZ, VX1
vfmul.d VX2, x4, VXS
vfmul.d VX3, x2, VXS
vfsub.d VX3, VXZ, VX3
vilvl.d x1, VX2 ,VX0
vilvh.d x2, VX2, VX0
vilvl.d x3, VX3 ,VX1
vilvh.d x4, VX3, VX1
vst x1, X, 0 * SIZE
vst x3, Y, 0 * SIZE
vst x2, X, 2 * SIZE
vst x4, Y, 2 * SIZE
addi.d X, X, 4 * SIZE
addi.d Y, Y, 4 * SIZE
#else
vld VX1, X, 4 * SIZE
vld VX3, Y, 4 * SIZE
vpickev.w x1, VX1, VX0
vpickod.w x2, VX1, VX0
vpickev.w x3, VX3, VX2
vpickod.w x4, VX3, VX2
vfmul.s VX0, x3, VXS
vfmul.s VX1, x1, VXS
vfsub.s VX1, VXZ, VX1
vfmul.s VX2, x4, VXS
vfmul.s VX3, x2, VXS
vfsub.s VX3, VXZ, VX3
vilvl.w x1, VX2 ,VX0
vilvh.w x2, VX2, VX0
vilvl.w x3, VX3 ,VX1
vilvh.w x4, VX3, VX1
vst x1, X, 0 * SIZE
vst x3, Y, 0 * SIZE
vst x2, X, 4 * SIZE
vst x4, Y, 4 * SIZE
addi.d X, X, 8 * SIZE
addi.d Y, Y, 8 * SIZE
#endif
addi.d I, I, -1
blt $r0, I, .L113
b .L997
.align 3

.L114: // C==0 S==0
vst VXZ, X, 0 * SIZE
vst VXZ, Y, 0 * SIZE
#ifdef DOUBLE
vst VXZ, X, 2 * SIZE
vst VXZ, Y, 2 * SIZE
addi.d X, X, 4 * SIZE
addi.d Y, Y, 4 * SIZE
#else
vst VXZ, X, 4 * SIZE
vst VXZ, Y, 4 * SIZE
addi.d X, X, 8 * SIZE
addi.d Y, Y, 8 * SIZE
#endif
addi.d I, I, -1
blt $r0, I, .L114
b .L997
.align 3

.L22:
#ifdef DOUBLE
srai.d I, N, 2
#endif
bge $r0, I, .L997
move YY, Y
move XX, X
CMPEQ $fcc0, C, a1
bcnez $fcc0, .L220
CMPEQ $fcc0, S, a1
bcnez $fcc0, .L222 // C!=0 S==0
b .L221 // C!=0 S!=0
.align 3

.L220:
CMPEQ $fcc0, S, a1
bcnez $fcc0, .L224 // C==0 S==0
b .L223 // C==0 S!=0
.align 3

.L221: // C!=0 S!=0
#ifdef DOUBLE
ld.d t1, X, 0 * SIZE
@@ -355,50 +198,6 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
vstelm.d VX1, YY, 0, 1
vstelm.d VX3, YY, 1 * SIZE, 1
add.d YY, YY, INCY

ld.d t1, X, 0 * SIZE
ld.d t2, X, 1 * SIZE
add.d X, X, INCX
ld.d t3, X, 0 * SIZE
ld.d t4, X, 1 * SIZE
vinsgr2vr.d x1, t1, 0
vinsgr2vr.d x2, t2, 0
vinsgr2vr.d x1, t3, 1
vinsgr2vr.d x2, t4, 1
add.d X, X, INCX
ld.d t1, Y, 0 * SIZE
ld.d t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.d t3, Y, 0 * SIZE
ld.d t4, Y, 1 * SIZE
vinsgr2vr.d x3, t1, 0
vinsgr2vr.d x4, t2, 0
vinsgr2vr.d x3, t3, 1
vinsgr2vr.d x4, t4, 1
add.d Y, Y, INCY
vfmul.d VX0, x1, VXC
vfmadd.d VX0, x3, VXS, VX0
vfmul.d VX1, x1, VXS
vfmsub.d VX1, x3, VXC, VX1
vfmul.d VX2, x2, VXC
vfmadd.d VX2, x4, VXS, VX2
vfmul.d VX3, x2, VXS
vfmsub.d VX3, x4, VXC, VX3
vstelm.d VX0, XX, 0, 0
vstelm.d VX2, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.d VX0, XX, 0, 1
vstelm.d VX2, XX, 1 * SIZE, 1
add.d XX, XX, INCX
vstelm.d VX1, YY, 0, 0
vstelm.d VX3, YY, 1 * SIZE, 0
add.d YY, YY, INCY
vstelm.d VX1, YY, 0, 1
vstelm.d VX3, YY, 1 * SIZE, 1
add.d YY, YY, INCY
addi.d I, I, -1
blt $r0, I, .L221
b .L995
#else
ld.w t1, X, 0 * SIZE
ld.w t2, X, 1 * SIZE
@@ -473,396 +272,12 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
vstelm.w VX1, YY, 0, 3
vstelm.w VX3, YY, 1 * SIZE, 3
add.d YY, YY, INCY
addi.d I, I, -1
blt $r0, I, .L221
b .L997
#endif
.align 3

.L222: // C!=0 S==0
#ifdef DOUBLE
ld.d t1, X, 0 * SIZE
ld.d t2, X, 1 * SIZE
add.d X, X, INCX
ld.d t3, X, 0 * SIZE
ld.d t4, X, 1 * SIZE
add.d X, X, INCX
vinsgr2vr.d x1, t1, 0
vinsgr2vr.d x2, t2, 0
vinsgr2vr.d x1, t3, 1
vinsgr2vr.d x2, t4, 1
ld.d t1, Y, 0 * SIZE
ld.d t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.d t3, Y, 0 * SIZE
ld.d t4, Y, 1 * SIZE
vinsgr2vr.d x3, t1, 0
vinsgr2vr.d x4, t2, 0
vinsgr2vr.d x3, t3, 1
vinsgr2vr.d x4, t4, 1
add.d Y, Y, INCY
vfmul.d VX0, x1, VXC
vfmul.d VX1, x3, VXC
vfmul.d VX2, x2, VXC
vfmul.d VX3, x4, VXC
vstelm.d VX0, XX, 0, 0
vstelm.d VX2, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.d VX0, XX, 0, 1
vstelm.d VX2, XX, 1 * SIZE, 1
add.d XX, XX, INCX
vstelm.d VX1, YY, 0, 0
vstelm.d VX3, YY, 1 * SIZE, 0
add.d YY, YY, INCY
vstelm.d VX1, YY, 0, 1
vstelm.d VX3, YY, 1 * SIZE, 1
add.d YY, YY, INCY

ld.d t1, X, 0 * SIZE
ld.d t2, X, 1 * SIZE
add.d X, X, INCX
ld.d t3, X, 0 * SIZE
ld.d t4, X, 1 * SIZE
vinsgr2vr.d x1, t1, 0
vinsgr2vr.d x2, t2, 0
vinsgr2vr.d x1, t3, 1
vinsgr2vr.d x2, t4, 1
add.d X, X, INCX
ld.d t1, Y, 0 * SIZE
ld.d t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.d t3, Y, 0 * SIZE
ld.d t4, Y, 1 * SIZE
vinsgr2vr.d x3, t1, 0
vinsgr2vr.d x4, t2, 0
vinsgr2vr.d x3, t3, 1
vinsgr2vr.d x4, t4, 1
add.d Y, Y, INCY
vfmul.d VX0, x1, VXC
vfmul.d VX1, x3, VXC
vfmul.d VX2, x2, VXC
vfmul.d VX3, x4, VXC
vstelm.d VX0, XX, 0, 0
vstelm.d VX2, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.d VX0, XX, 0, 1
vstelm.d VX2, XX, 1 * SIZE, 1
add.d XX, XX, INCX
vstelm.d VX1, YY, 0, 0
vstelm.d VX3, YY, 1 * SIZE, 0
add.d YY, YY, INCY
vstelm.d VX1, YY, 0, 1
vstelm.d VX3, YY, 1 * SIZE, 1
add.d YY, YY, INCY
addi.d I, I, -1
blt $r0, I, .L222
b .L995
#else
ld.w t1, X, 0 * SIZE
ld.w t2, X, 1 * SIZE
add.d X, X, INCX
ld.w t3, X, 0 * SIZE
ld.w t4, X, 1 * SIZE
add.d X, X, INCX
vinsgr2vr.w x1, t1, 0
vinsgr2vr.w x2, t2, 0
vinsgr2vr.w x1, t3, 1
vinsgr2vr.w x2, t4, 1
ld.w t1, Y, 0 * SIZE
ld.w t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.w t3, Y, 0 * SIZE
ld.w t4, Y, 1 * SIZE
add.d Y, Y, INCY
vinsgr2vr.w x3, t1, 0
vinsgr2vr.w x4, t2, 0
vinsgr2vr.w x3, t3, 1
vinsgr2vr.w x4, t4, 1
ld.w t1, X, 0 * SIZE
ld.w t2, X, 1 * SIZE
add.d X, X, INCX
ld.w t3, X, 0 * SIZE
ld.w t4, X, 1 * SIZE
vinsgr2vr.w x1, t1, 2
vinsgr2vr.w x2, t2, 2
vinsgr2vr.w x1, t3, 3
vinsgr2vr.w x2, t4, 3
add.d X, X, INCX
ld.w t1, Y, 0 * SIZE
ld.w t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.w t3, Y, 0 * SIZE
ld.w t4, Y, 1 * SIZE
vinsgr2vr.w x3, t1, 2
vinsgr2vr.w x4, t2, 2
vinsgr2vr.w x3, t3, 3
vinsgr2vr.w x4, t4, 3
add.d Y, Y, INCY
vfmul.s VX0, x1, VXC
vfmul.s VX1, x3, VXC
vfmul.s VX2, x2, VXC
vfmul.s VX3, x4, VXC
vstelm.w VX0, XX, 0, 0
vstelm.w VX2, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.w VX0, XX, 0, 1
vstelm.w VX2, XX, 1 * SIZE, 1
add.d XX, XX, INCX
vstelm.w VX0, XX, 0, 2
vstelm.w VX2, XX, 1 * SIZE, 2
add.d XX, XX, INCX
vstelm.w VX0, XX, 0, 3
vstelm.w VX2, XX, 1 * SIZE, 3
add.d XX, XX, INCX
vstelm.w VX1, YY, 0, 0
vstelm.w VX3, YY, 1 * SIZE, 0
add.d YY, YY, INCY
vstelm.w VX1, YY, 0, 1
vstelm.w VX3, YY, 1 * SIZE, 1
add.d YY, YY, INCY
vstelm.w VX1, YY, 0, 2
vstelm.w VX3, YY, 1 * SIZE, 2
add.d YY, YY, INCY
vstelm.w VX1, YY, 0, 3
vstelm.w VX3, YY, 1 * SIZE, 3
add.d YY, YY, INCY
addi.d I, I, -1
blt $r0, I, .L222
b .L997
#endif
.align 3

.L223: // C==0 S!=0
#ifdef DOUBLE
ld.d t1, X, 0 * SIZE
ld.d t2, X, 1 * SIZE
add.d X, X, INCX
ld.d t3, X, 0 * SIZE
ld.d t4, X, 1 * SIZE
add.d X, X, INCX
vinsgr2vr.d x1, t1, 0
vinsgr2vr.d x2, t2, 0
vinsgr2vr.d x1, t3, 1
vinsgr2vr.d x2, t4, 1
ld.d t1, Y, 0 * SIZE
ld.d t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.d t3, Y, 0 * SIZE
ld.d t4, Y, 1 * SIZE
vinsgr2vr.d x3, t1, 0
vinsgr2vr.d x4, t2, 0
vinsgr2vr.d x3, t3, 1
vinsgr2vr.d x4, t4, 1
add.d Y, Y, INCY
vfmul.d VX0, x3, VXS
vfmul.d VX1, x1, VXS
vfsub.d VX1, VXZ, VX1
vfmul.d VX2, x4, VXS
vfmul.d VX3, x2, VXS
vfsub.d VX3, VXZ, VX3
vstelm.d VX0, XX, 0, 0
vstelm.d VX2, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.d VX0, XX, 0, 1
vstelm.d VX2, XX, 1 * SIZE, 1
add.d XX, XX, INCX
vstelm.d VX1, YY, 0, 0
vstelm.d VX3, YY, 1 * SIZE, 0
add.d YY, YY, INCY
vstelm.d VX1, YY, 0, 1
vstelm.d VX3, YY, 1 * SIZE, 1
add.d YY, YY, INCY

ld.d t1, X, 0 * SIZE
ld.d t2, X, 1 * SIZE
add.d X, X, INCX
ld.d t3, X, 0 * SIZE
ld.d t4, X, 1 * SIZE
vinsgr2vr.d x1, t1, 0
vinsgr2vr.d x2, t2, 0
vinsgr2vr.d x1, t3, 1
vinsgr2vr.d x2, t4, 1
add.d X, X, INCX
ld.d t1, Y, 0 * SIZE
ld.d t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.d t3, Y, 0 * SIZE
ld.d t4, Y, 1 * SIZE
vinsgr2vr.d x3, t1, 0
vinsgr2vr.d x4, t2, 0
vinsgr2vr.d x3, t3, 1
vinsgr2vr.d x4, t4, 1
add.d Y, Y, INCY
vfmul.d VX0, x3, VXS
vfmul.d VX1, x1, VXS
vfsub.d VX1, VXZ, VX1
vfmul.d VX2, x4, VXS
vfmul.d VX3, x2, VXS
vfsub.d VX3, VXZ, VX3
vstelm.d VX0, XX, 0, 0
vstelm.d VX2, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.d VX0, XX, 0, 1
vstelm.d VX2, XX, 1 * SIZE, 1
add.d XX, XX, INCX
vstelm.d VX1, YY, 0, 0
vstelm.d VX3, YY, 1 * SIZE, 0
add.d YY, YY, INCY
vstelm.d VX1, YY, 0, 1
vstelm.d VX3, YY, 1 * SIZE, 1
add.d YY, YY, INCY
addi.d I, I, -1
blt $r0, I, .L223
b .L995
#else
ld.w t1, X, 0 * SIZE
ld.w t2, X, 1 * SIZE
add.d X, X, INCX
ld.w t3, X, 0 * SIZE
ld.w t4, X, 1 * SIZE
add.d X, X, INCX
vinsgr2vr.w x1, t1, 0
vinsgr2vr.w x2, t2, 0
vinsgr2vr.w x1, t3, 1
vinsgr2vr.w x2, t4, 1
ld.w t1, Y, 0 * SIZE
ld.w t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.w t3, Y, 0 * SIZE
ld.w t4, Y, 1 * SIZE
add.d Y, Y, INCY
vinsgr2vr.w x3, t1, 0
vinsgr2vr.w x4, t2, 0
vinsgr2vr.w x3, t3, 1
vinsgr2vr.w x4, t4, 1
ld.w t1, X, 0 * SIZE
ld.w t2, X, 1 * SIZE
add.d X, X, INCX
ld.w t3, X, 0 * SIZE
ld.w t4, X, 1 * SIZE
vinsgr2vr.w x1, t1, 2
vinsgr2vr.w x2, t2, 2
vinsgr2vr.w x1, t3, 3
vinsgr2vr.w x2, t4, 3
add.d X, X, INCX
ld.w t1, Y, 0 * SIZE
ld.w t2, Y, 1 * SIZE
add.d Y, Y, INCY
ld.w t3, Y, 0 * SIZE
ld.w t4, Y, 1 * SIZE
vinsgr2vr.w x3, t1, 2
vinsgr2vr.w x4, t2, 2
vinsgr2vr.w x3, t3, 3
vinsgr2vr.w x4, t4, 3
add.d Y, Y, INCY
vfmul.s VX0, x3, VXS
vfmul.s VX1, x1, VXS
vfsub.s VX1, VXZ, VX1
vfmul.s VX2, x4, VXS
vfmul.s VX3, x2, VXS
vfsub.s VX3, VXZ, VX3
vstelm.w VX0, XX, 0, 0
vstelm.w VX2, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.w VX0, XX, 0, 1
vstelm.w VX2, XX, 1 * SIZE, 1
add.d XX, XX, INCX
vstelm.w VX0, XX, 0, 2
vstelm.w VX2, XX, 1 * SIZE, 2
add.d XX, XX, INCX
vstelm.w VX0, XX, 0, 3
vstelm.w VX2, XX, 1 * SIZE, 3
add.d XX, XX, INCX
vstelm.w VX1, YY, 0, 0
vstelm.w VX3, YY, 1 * SIZE, 0
add.d YY, YY, INCY
vstelm.w VX1, YY, 0, 1
vstelm.w VX3, YY, 1 * SIZE, 1
add.d YY, YY, INCY
vstelm.w VX1, YY, 0, 2
vstelm.w VX3, YY, 1 * SIZE, 2
add.d YY, YY, INCY
vstelm.w VX1, YY, 0, 3
vstelm.w VX3, YY, 1 * SIZE, 3
add.d YY, YY, INCY
addi.d I, I, -1
blt $r0, I, .L223
b .L997
#endif
.align 3

.L224: // C==0 S==0
#ifdef DOUBLE
vstelm.d VXZ, XX, 0, 0
vstelm.d VXZ, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.d VXZ, XX, 0, 0
vstelm.d VXZ, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.d VXZ, XX, 0, 0
vstelm.d VXZ, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.d VXZ, XX, 0, 0
vstelm.d VXZ, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.d VXZ, YY, 0, 0
vstelm.d VXZ, YY, 1 * SIZE, 0
add.d YY, YY, INCY
vstelm.d VXZ, YY, 0, 0
vstelm.d VXZ, YY, 1 * SIZE, 0
add.d YY, YY, INCY
vstelm.d VXZ, YY, 0, 0
vstelm.d VXZ, YY, 1 * SIZE, 0
add.d YY, YY, INCY
vstelm.d VXZ, YY, 0, 0
vstelm.d VXZ, YY, 1 * SIZE, 0
add.d YY, YY, INCY
addi.d I, I, -1
blt $r0, I, .L224
move X, XX
move Y, YY
b .L995
#else
vstelm.w VXZ, XX, 0, 0
vstelm.w VXZ, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.w VXZ, XX, 0, 0
vstelm.w VXZ, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.w VXZ, XX, 0, 0
vstelm.w VXZ, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.w VXZ, XX, 0, 0
vstelm.w VXZ, XX, 1 * SIZE, 0
add.d XX, XX, INCX
vstelm.w VXZ, YY, 0, 0
vstelm.w VXZ, YY, 1 * SIZE, 0
add.d YY, YY, INCY
vstelm.w VXZ, YY, 0, 0
vstelm.w VXZ, YY, 1 * SIZE, 0
add.d YY, YY, INCY
vstelm.w VXZ, YY, 0, 0
vstelm.w VXZ, YY, 1 * SIZE, 0
add.d YY, YY, INCY
vstelm.w VXZ, YY, 0, 0
vstelm.w VXZ, YY, 1 * SIZE, 0
add.d YY, YY, INCY
addi.d I, I, -1
blt $r0, I, .L224
move X, XX
move Y, YY
blt $r0, I, .L221
b .L997
#endif
.align 3

#ifdef DOUBLE
.L995:
andi I, N, 3
bge $r0, I, .L999
b .L998
.align 3

#endif
.L996:
move I, N
b .L998


+ 113
- 1292
kernel/loongarch64/rot_lsx.S
File diff suppressed because it is too large
View File


Write Preview
Loading…
Cancel
Save