Welcome to mirror list, hosted at ThFree Co, Russian Federation.

github.com/FFmpeg/FFmpeg.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
path: root/libavcodec/ppc/hpeldsp_altivec.c
diff options
context:
space:
mode:
Diffstat (limited to 'libavcodec/ppc/hpeldsp_altivec.c')
-rw-r--r--libavcodec/ppc/hpeldsp_altivec.c233
1 files changed, 77 insertions, 156 deletions
diff --git a/libavcodec/ppc/hpeldsp_altivec.c b/libavcodec/ppc/hpeldsp_altivec.c
index 780240ba20..87a1f05b6a 100644
--- a/libavcodec/ppc/hpeldsp_altivec.c
+++ b/libavcodec/ppc/hpeldsp_altivec.c
@@ -3,20 +3,20 @@
* Copyright (c) 2002 Dieter Shirley
* Copyright (c) 2003-2004 Romain Dolbeau <romain@dolbeau.org>
*
- * This file is part of Libav.
+ * This file is part of FFmpeg.
*
- * Libav is free software; you can redistribute it and/or
+ * FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
- * Libav is distributed in the hope that it will be useful,
+ * FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
- * License along with Libav; if not, write to the Free Software
+ * License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
@@ -34,16 +34,15 @@
#include "libavcodec/hpeldsp.h"
#include "hpeldsp_altivec.h"
-#if HAVE_ALTIVEC && HAVE_BIGENDIAN
+#if HAVE_ALTIVEC
/* next one assumes that ((line_size % 16) == 0) */
void ff_put_pixels16_altivec(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
{
- register vector unsigned char pixelsv1, pixelsv2;
- register vector unsigned char pixelsv1B, pixelsv2B;
- register vector unsigned char pixelsv1C, pixelsv2C;
- register vector unsigned char pixelsv1D, pixelsv2D;
+ register vector unsigned char pixelsv1;
+ register vector unsigned char pixelsv1B;
+ register vector unsigned char pixelsv1C;
+ register vector unsigned char pixelsv1D;
- register vector unsigned char perm = vec_lvsl(0, pixels);
int i;
register ptrdiff_t line_size_2 = line_size << 1;
register ptrdiff_t line_size_3 = line_size + line_size_2;
@@ -55,22 +54,14 @@ void ff_put_pixels16_altivec(uint8_t *block, const uint8_t *pixels, ptrdiff_t li
// -funroll-loops w/ this is bad - 74 cycles again.
// all this is on a 7450, tuning for the 7450
for (i = 0; i < h; i += 4) {
- pixelsv1 = vec_ld( 0, pixels);
- pixelsv2 = vec_ld(15, pixels);
- pixelsv1B = vec_ld(line_size, pixels);
- pixelsv2B = vec_ld(15 + line_size, pixels);
- pixelsv1C = vec_ld(line_size_2, pixels);
- pixelsv2C = vec_ld(15 + line_size_2, pixels);
- pixelsv1D = vec_ld(line_size_3, pixels);
- pixelsv2D = vec_ld(15 + line_size_3, pixels);
- vec_st(vec_perm(pixelsv1, pixelsv2, perm),
- 0, (unsigned char*)block);
- vec_st(vec_perm(pixelsv1B, pixelsv2B, perm),
- line_size, (unsigned char*)block);
- vec_st(vec_perm(pixelsv1C, pixelsv2C, perm),
- line_size_2, (unsigned char*)block);
- vec_st(vec_perm(pixelsv1D, pixelsv2D, perm),
- line_size_3, (unsigned char*)block);
+ pixelsv1 = unaligned_load( 0, pixels);
+ pixelsv1B = unaligned_load(line_size, pixels);
+ pixelsv1C = unaligned_load(line_size_2, pixels);
+ pixelsv1D = unaligned_load(line_size_3, pixels);
+ VEC_ST(pixelsv1, 0, (unsigned char*)block);
+ VEC_ST(pixelsv1B, line_size, (unsigned char*)block);
+ VEC_ST(pixelsv1C, line_size_2, (unsigned char*)block);
+ VEC_ST(pixelsv1D, line_size_3, (unsigned char*)block);
pixels+=line_size_4;
block +=line_size_4;
}
@@ -80,15 +71,12 @@ void ff_put_pixels16_altivec(uint8_t *block, const uint8_t *pixels, ptrdiff_t li
#define op_avg(a,b) a = ( ((a)|(b)) - ((((a)^(b))&0xFEFEFEFEUL)>>1) )
void ff_avg_pixels16_altivec(uint8_t *block, const uint8_t *pixels, ptrdiff_t line_size, int h)
{
- register vector unsigned char pixelsv1, pixelsv2, pixelsv, blockv;
- register vector unsigned char perm = vec_lvsl(0, pixels);
- int i;
+ register vector unsigned char pixelsv, blockv;
+ int i;
for (i = 0; i < h; i++) {
- pixelsv1 = vec_ld( 0, pixels);
- pixelsv2 = vec_ld(16,pixels);
blockv = vec_ld(0, block);
- pixelsv = vec_perm(pixelsv1, pixelsv2, perm);
+ pixelsv = VEC_LD( 0, pixels);
blockv = vec_avg(blockv,pixelsv);
vec_st(blockv, 0, (unsigned char*)block);
pixels+=line_size;
@@ -108,9 +96,7 @@ static void avg_pixels8_altivec(uint8_t * block, const uint8_t * pixels, ptrdiff
int rightside = ((unsigned long)block & 0x0000000F);
blockv = vec_ld(0, block);
- pixelsv1 = vec_ld( 0, pixels);
- pixelsv2 = vec_ld(16, pixels);
- pixelsv = vec_perm(pixelsv1, pixelsv2, vec_lvsl(0, pixels));
+ pixelsv = VEC_LD( 0, pixels);
if (rightside) {
pixelsv = vec_perm(blockv, pixelsv, vcprm(0,1,s0,s1));
@@ -132,21 +118,16 @@ static void put_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels, ptrdi
{
register int i;
register vector unsigned char pixelsv1, pixelsv2, pixelsavg;
- register vector unsigned char blockv, temp1, temp2;
+ register vector unsigned char blockv;
register vector unsigned short pixelssum1, pixelssum2, temp3;
register const vector unsigned char vczero = (const vector unsigned char)vec_splat_u8(0);
register const vector unsigned short vctwo = (const vector unsigned short)vec_splat_u16(2);
- temp1 = vec_ld(0, pixels);
- temp2 = vec_ld(16, pixels);
- pixelsv1 = vec_perm(temp1, temp2, vec_lvsl(0, pixels));
- if ((((unsigned long)pixels) & 0x0000000F) == 0x0000000F) {
- pixelsv2 = temp2;
- } else {
- pixelsv2 = vec_perm(temp1, temp2, vec_lvsl(1, pixels));
- }
- pixelsv1 = vec_mergeh(vczero, pixelsv1);
- pixelsv2 = vec_mergeh(vczero, pixelsv2);
+ pixelsv1 = VEC_LD(0, pixels);
+ pixelsv2 = VEC_LD(1, pixels);
+ pixelsv1 = VEC_MERGEH(vczero, pixelsv1);
+ pixelsv2 = VEC_MERGEH(vczero, pixelsv2);
+
pixelssum1 = vec_add((vector unsigned short)pixelsv1,
(vector unsigned short)pixelsv2);
pixelssum1 = vec_add(pixelssum1, vctwo);
@@ -155,17 +136,10 @@ static void put_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels, ptrdi
int rightside = ((unsigned long)block & 0x0000000F);
blockv = vec_ld(0, block);
- temp1 = vec_ld(line_size, pixels);
- temp2 = vec_ld(line_size + 16, pixels);
- pixelsv1 = vec_perm(temp1, temp2, vec_lvsl(line_size, pixels));
- if (((((unsigned long)pixels) + line_size) & 0x0000000F) == 0x0000000F) {
- pixelsv2 = temp2;
- } else {
- pixelsv2 = vec_perm(temp1, temp2, vec_lvsl(line_size + 1, pixels));
- }
-
- pixelsv1 = vec_mergeh(vczero, pixelsv1);
- pixelsv2 = vec_mergeh(vczero, pixelsv2);
+ pixelsv1 = unaligned_load(line_size, pixels);
+ pixelsv2 = unaligned_load(line_size+1, pixels);
+ pixelsv1 = VEC_MERGEH(vczero, pixelsv1);
+ pixelsv2 = VEC_MERGEH(vczero, pixelsv2);
pixelssum2 = vec_add((vector unsigned short)pixelsv1,
(vector unsigned short)pixelsv2);
temp3 = vec_add(pixelssum1, pixelssum2);
@@ -191,22 +165,16 @@ static void put_no_rnd_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels
{
register int i;
register vector unsigned char pixelsv1, pixelsv2, pixelsavg;
- register vector unsigned char blockv, temp1, temp2;
+ register vector unsigned char blockv;
register vector unsigned short pixelssum1, pixelssum2, temp3;
register const vector unsigned char vczero = (const vector unsigned char)vec_splat_u8(0);
register const vector unsigned short vcone = (const vector unsigned short)vec_splat_u16(1);
register const vector unsigned short vctwo = (const vector unsigned short)vec_splat_u16(2);
- temp1 = vec_ld(0, pixels);
- temp2 = vec_ld(16, pixels);
- pixelsv1 = vec_perm(temp1, temp2, vec_lvsl(0, pixels));
- if ((((unsigned long)pixels) & 0x0000000F) == 0x0000000F) {
- pixelsv2 = temp2;
- } else {
- pixelsv2 = vec_perm(temp1, temp2, vec_lvsl(1, pixels));
- }
- pixelsv1 = vec_mergeh(vczero, pixelsv1);
- pixelsv2 = vec_mergeh(vczero, pixelsv2);
+ pixelsv1 = VEC_LD(0, pixels);
+ pixelsv2 = VEC_LD(1, pixels);
+ pixelsv1 = VEC_MERGEH(vczero, pixelsv1);
+ pixelsv2 = VEC_MERGEH(vczero, pixelsv2);
pixelssum1 = vec_add((vector unsigned short)pixelsv1,
(vector unsigned short)pixelsv2);
pixelssum1 = vec_add(pixelssum1, vcone);
@@ -215,17 +183,10 @@ static void put_no_rnd_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels
int rightside = ((unsigned long)block & 0x0000000F);
blockv = vec_ld(0, block);
- temp1 = vec_ld(line_size, pixels);
- temp2 = vec_ld(line_size + 16, pixels);
- pixelsv1 = vec_perm(temp1, temp2, vec_lvsl(line_size, pixels));
- if (((((unsigned long)pixels) + line_size) & 0x0000000F) == 0x0000000F) {
- pixelsv2 = temp2;
- } else {
- pixelsv2 = vec_perm(temp1, temp2, vec_lvsl(line_size + 1, pixels));
- }
-
- pixelsv1 = vec_mergeh(vczero, pixelsv1);
- pixelsv2 = vec_mergeh(vczero, pixelsv2);
+ pixelsv1 = unaligned_load(line_size, pixels);
+ pixelsv2 = unaligned_load(line_size+1, pixels);
+ pixelsv1 = VEC_MERGEH(vczero, pixelsv1);
+ pixelsv2 = VEC_MERGEH(vczero, pixelsv2);
pixelssum2 = vec_add((vector unsigned short)pixelsv1,
(vector unsigned short)pixelsv2);
temp3 = vec_add(pixelssum1, pixelssum2);
@@ -251,24 +212,18 @@ static void put_pixels16_xy2_altivec(uint8_t * block, const uint8_t * pixels, pt
{
register int i;
register vector unsigned char pixelsv1, pixelsv2, pixelsv3, pixelsv4;
- register vector unsigned char blockv, temp1, temp2;
+ register vector unsigned char blockv;
register vector unsigned short temp3, temp4,
pixelssum1, pixelssum2, pixelssum3, pixelssum4;
register const vector unsigned char vczero = (const vector unsigned char)vec_splat_u8(0);
register const vector unsigned short vctwo = (const vector unsigned short)vec_splat_u16(2);
- temp1 = vec_ld(0, pixels);
- temp2 = vec_ld(16, pixels);
- pixelsv1 = vec_perm(temp1, temp2, vec_lvsl(0, pixels));
- if ((((unsigned long)pixels) & 0x0000000F) == 0x0000000F) {
- pixelsv2 = temp2;
- } else {
- pixelsv2 = vec_perm(temp1, temp2, vec_lvsl(1, pixels));
- }
- pixelsv3 = vec_mergel(vczero, pixelsv1);
- pixelsv4 = vec_mergel(vczero, pixelsv2);
- pixelsv1 = vec_mergeh(vczero, pixelsv1);
- pixelsv2 = vec_mergeh(vczero, pixelsv2);
+ pixelsv1 = VEC_LD(0, pixels);
+ pixelsv2 = VEC_LD(1, pixels);
+ pixelsv3 = VEC_MERGEL(vczero, pixelsv1);
+ pixelsv4 = VEC_MERGEL(vczero, pixelsv2);
+ pixelsv1 = VEC_MERGEH(vczero, pixelsv1);
+ pixelsv2 = VEC_MERGEH(vczero, pixelsv2);
pixelssum3 = vec_add((vector unsigned short)pixelsv3,
(vector unsigned short)pixelsv4);
pixelssum3 = vec_add(pixelssum3, vctwo);
@@ -279,20 +234,13 @@ static void put_pixels16_xy2_altivec(uint8_t * block, const uint8_t * pixels, pt
for (i = 0; i < h ; i++) {
blockv = vec_ld(0, block);
- temp1 = vec_ld(line_size, pixels);
- temp2 = vec_ld(line_size + 16, pixels);
- pixelsv1 = vec_perm(temp1, temp2, vec_lvsl(line_size, pixels));
- if (((((unsigned long)pixels) + line_size) & 0x0000000F) == 0x0000000F) {
- pixelsv2 = temp2;
- } else {
- pixelsv2 = vec_perm(temp1, temp2, vec_lvsl(line_size + 1, pixels));
- }
-
- pixelsv3 = vec_mergel(vczero, pixelsv1);
- pixelsv4 = vec_mergel(vczero, pixelsv2);
- pixelsv1 = vec_mergeh(vczero, pixelsv1);
- pixelsv2 = vec_mergeh(vczero, pixelsv2);
+ pixelsv1 = unaligned_load(line_size, pixels);
+ pixelsv2 = unaligned_load(line_size+1, pixels);
+ pixelsv3 = VEC_MERGEL(vczero, pixelsv1);
+ pixelsv4 = VEC_MERGEL(vczero, pixelsv2);
+ pixelsv1 = VEC_MERGEH(vczero, pixelsv1);
+ pixelsv2 = VEC_MERGEH(vczero, pixelsv2);
pixelssum4 = vec_add((vector unsigned short)pixelsv3,
(vector unsigned short)pixelsv4);
pixelssum2 = vec_add((vector unsigned short)pixelsv1,
@@ -319,25 +267,19 @@ static void put_no_rnd_pixels16_xy2_altivec(uint8_t * block, const uint8_t * pix
{
register int i;
register vector unsigned char pixelsv1, pixelsv2, pixelsv3, pixelsv4;
- register vector unsigned char blockv, temp1, temp2;
+ register vector unsigned char blockv;
register vector unsigned short temp3, temp4,
pixelssum1, pixelssum2, pixelssum3, pixelssum4;
register const vector unsigned char vczero = (const vector unsigned char)vec_splat_u8(0);
register const vector unsigned short vcone = (const vector unsigned short)vec_splat_u16(1);
register const vector unsigned short vctwo = (const vector unsigned short)vec_splat_u16(2);
- temp1 = vec_ld(0, pixels);
- temp2 = vec_ld(16, pixels);
- pixelsv1 = vec_perm(temp1, temp2, vec_lvsl(0, pixels));
- if ((((unsigned long)pixels) & 0x0000000F) == 0x0000000F) {
- pixelsv2 = temp2;
- } else {
- pixelsv2 = vec_perm(temp1, temp2, vec_lvsl(1, pixels));
- }
- pixelsv3 = vec_mergel(vczero, pixelsv1);
- pixelsv4 = vec_mergel(vczero, pixelsv2);
- pixelsv1 = vec_mergeh(vczero, pixelsv1);
- pixelsv2 = vec_mergeh(vczero, pixelsv2);
+ pixelsv1 = VEC_LD(0, pixels);
+ pixelsv2 = VEC_LD(1, pixels);
+ pixelsv3 = VEC_MERGEL(vczero, pixelsv1);
+ pixelsv4 = VEC_MERGEL(vczero, pixelsv2);
+ pixelsv1 = VEC_MERGEH(vczero, pixelsv1);
+ pixelsv2 = VEC_MERGEH(vczero, pixelsv2);
pixelssum3 = vec_add((vector unsigned short)pixelsv3,
(vector unsigned short)pixelsv4);
pixelssum3 = vec_add(pixelssum3, vcone);
@@ -346,22 +288,13 @@ static void put_no_rnd_pixels16_xy2_altivec(uint8_t * block, const uint8_t * pix
pixelssum1 = vec_add(pixelssum1, vcone);
for (i = 0; i < h ; i++) {
- blockv = vec_ld(0, block);
-
- temp1 = vec_ld(line_size, pixels);
- temp2 = vec_ld(line_size + 16, pixels);
- pixelsv1 = vec_perm(temp1, temp2, vec_lvsl(line_size, pixels));
- if (((((unsigned long)pixels) + line_size) & 0x0000000F) == 0x0000000F) {
- pixelsv2 = temp2;
- } else {
- pixelsv2 = vec_perm(temp1, temp2, vec_lvsl(line_size + 1, pixels));
- }
-
- pixelsv3 = vec_mergel(vczero, pixelsv1);
- pixelsv4 = vec_mergel(vczero, pixelsv2);
- pixelsv1 = vec_mergeh(vczero, pixelsv1);
- pixelsv2 = vec_mergeh(vczero, pixelsv2);
+ pixelsv1 = unaligned_load(line_size, pixels);
+ pixelsv2 = unaligned_load(line_size+1, pixels);
+ pixelsv3 = VEC_MERGEL(vczero, pixelsv1);
+ pixelsv4 = VEC_MERGEL(vczero, pixelsv2);
+ pixelsv1 = VEC_MERGEH(vczero, pixelsv1);
+ pixelsv2 = VEC_MERGEH(vczero, pixelsv2);
pixelssum4 = vec_add((vector unsigned short)pixelsv3,
(vector unsigned short)pixelsv4);
pixelssum2 = vec_add((vector unsigned short)pixelsv1,
@@ -376,7 +309,7 @@ static void put_no_rnd_pixels16_xy2_altivec(uint8_t * block, const uint8_t * pix
blockv = vec_packsu(temp3, temp4);
- vec_st(blockv, 0, block);
+ VEC_ST(blockv, 0, block);
block += line_size;
pixels += line_size;
@@ -388,7 +321,7 @@ static void avg_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels, ptrdi
{
register int i;
register vector unsigned char pixelsv1, pixelsv2, pixelsavg;
- register vector unsigned char blockv, temp1, temp2, blocktemp;
+ register vector unsigned char blockv, blocktemp;
register vector unsigned short pixelssum1, pixelssum2, temp3;
register const vector unsigned char vczero = (const vector unsigned char)
@@ -396,16 +329,10 @@ static void avg_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels, ptrdi
register const vector unsigned short vctwo = (const vector unsigned short)
vec_splat_u16(2);
- temp1 = vec_ld(0, pixels);
- temp2 = vec_ld(16, pixels);
- pixelsv1 = vec_perm(temp1, temp2, vec_lvsl(0, pixels));
- if ((((unsigned long)pixels) & 0x0000000F) == 0x0000000F) {
- pixelsv2 = temp2;
- } else {
- pixelsv2 = vec_perm(temp1, temp2, vec_lvsl(1, pixels));
- }
- pixelsv1 = vec_mergeh(vczero, pixelsv1);
- pixelsv2 = vec_mergeh(vczero, pixelsv2);
+ pixelsv1 = VEC_LD(0, pixels);
+ pixelsv2 = VEC_LD(1, pixels);
+ pixelsv1 = VEC_MERGEH(vczero, pixelsv1);
+ pixelsv2 = VEC_MERGEH(vczero, pixelsv2);
pixelssum1 = vec_add((vector unsigned short)pixelsv1,
(vector unsigned short)pixelsv2);
pixelssum1 = vec_add(pixelssum1, vctwo);
@@ -414,17 +341,11 @@ static void avg_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels, ptrdi
int rightside = ((unsigned long)block & 0x0000000F);
blockv = vec_ld(0, block);
- temp1 = vec_ld(line_size, pixels);
- temp2 = vec_ld(line_size + 16, pixels);
- pixelsv1 = vec_perm(temp1, temp2, vec_lvsl(line_size, pixels));
- if (((((unsigned long)pixels) + line_size) & 0x0000000F) == 0x0000000F) {
- pixelsv2 = temp2;
- } else {
- pixelsv2 = vec_perm(temp1, temp2, vec_lvsl(line_size + 1, pixels));
- }
+ pixelsv1 = unaligned_load(line_size, pixels);
+ pixelsv2 = unaligned_load(line_size+1, pixels);
- pixelsv1 = vec_mergeh(vczero, pixelsv1);
- pixelsv2 = vec_mergeh(vczero, pixelsv2);
+ pixelsv1 = VEC_MERGEH(vczero, pixelsv1);
+ pixelsv2 = VEC_MERGEH(vczero, pixelsv2);
pixelssum2 = vec_add((vector unsigned short)pixelsv1,
(vector unsigned short)pixelsv2);
temp3 = vec_add(pixelssum1, pixelssum2);
@@ -449,7 +370,7 @@ static void avg_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels, ptrdi
av_cold void ff_hpeldsp_init_ppc(HpelDSPContext *c, int flags)
{
-#if HAVE_ALTIVEC && HAVE_BIGENDIAN
+#if HAVE_ALTIVEC
if (!PPC_ALTIVEC(av_get_cpu_flags()))
return;
generated by cgit v1.2.3 (git 2.25.1) at 2024-07-10 23:06:52 +0300