Mercurial > libavcodec.hg

changeset 3554:ce5554dd79ce libavcodec

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Cosmetics: 2->4 spaces and some braces
author lu_zero
date Sun, 06 Aug 2006 23:15:32 +0000
parents a542b0325239
children 5ea82888103e
files ppc/dsputil_altivec.c
diffstat 1 files changed, 121 insertions(+), 130 deletions(-) [+]
line wrap: on
line diff
--- a/ppc/dsputil_altivec.c	Sun Aug 06 03:44:31 2006 +0000
+++ b/ppc/dsputil_altivec.c	Sun Aug 06 23:15:32 2006 +0000
@@ -1310,29 +1310,28 @@
 
 int hadamard8_diff8x8_altivec(/*MpegEncContext*/ void *s, uint8_t *dst, uint8_t *src, int stride, int h){
 POWERPC_PERF_DECLARE(altivec_hadamard8_diff8x8_num, 1);
-  int sum;
-  register const_vector unsigned char vzero = (const_vector unsigned char)vec_splat_u8(0);
-  register vector signed short temp0, temp1, temp2, temp3, temp4, temp5, temp6, temp7;
+    int sum;
+    register const_vector unsigned char vzero =
+                            (const_vector unsigned char)vec_splat_u8(0);
+    register vector signed short temp0, temp1, temp2, temp3, temp4,
+                                 temp5, temp6, temp7;
 POWERPC_PERF_START_COUNT(altivec_hadamard8_diff8x8_num, 1);
   {
-    register const_vector signed short vprod1 = (const_vector signed short)AVV( 1,-1, 1,-1, 1,-1, 1,-1);
-    register const_vector signed short vprod2 = (const_vector signed short)AVV( 1, 1,-1,-1, 1, 1,-1,-1);
-    register const_vector signed short vprod3 = (const_vector signed short)AVV( 1, 1, 1, 1,-1,-1,-1,-1);
+    register const_vector signed short vprod1 =(const_vector signed short)
+                                        AVV( 1,-1, 1,-1, 1,-1, 1,-1);
+    register const_vector signed short vprod2 =(const_vector signed short)
+                                        AVV( 1, 1,-1,-1, 1, 1,-1,-1);
+    register const_vector signed short vprod3 =(const_vector signed short)
+                                        AVV( 1, 1, 1, 1,-1,-1,-1,-1);
     register const_vector unsigned char perm1 = (const_vector unsigned char)
-      AVV(0x02, 0x03, 0x00, 0x01,
-       0x06, 0x07, 0x04, 0x05,
-       0x0A, 0x0B, 0x08, 0x09,
-       0x0E, 0x0F, 0x0C, 0x0D);
+      AVV(0x02, 0x03, 0x00, 0x01, 0x06, 0x07, 0x04, 0x05,
+          0x0A, 0x0B, 0x08, 0x09, 0x0E, 0x0F, 0x0C, 0x0D);
     register const_vector unsigned char perm2 = (const_vector unsigned char)
-      AVV(0x04, 0x05, 0x06, 0x07,
-       0x00, 0x01, 0x02, 0x03,
-       0x0C, 0x0D, 0x0E, 0x0F,
-       0x08, 0x09, 0x0A, 0x0B);
+      AVV(0x04, 0x05, 0x06, 0x07, 0x00, 0x01, 0x02, 0x03,
+          0x0C, 0x0D, 0x0E, 0x0F, 0x08, 0x09, 0x0A, 0x0B);
     register const_vector unsigned char perm3 = (const_vector unsigned char)
-      AVV(0x08, 0x09, 0x0A, 0x0B,
-       0x0C, 0x0D, 0x0E, 0x0F,
-       0x00, 0x01, 0x02, 0x03,
-       0x04, 0x05, 0x06, 0x07);
+      AVV(0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
+          0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07);
 
 #define ONEITERBUTTERFLY(i, res)                                        \
     {                                                                   \
@@ -1443,45 +1442,46 @@
 */
 
 static int hadamard8_diff16x8_altivec(/*MpegEncContext*/ void *s, uint8_t *dst, uint8_t *src, int stride, int h) {
-  int sum;
-  register vector signed short
-    temp0 REG_v(v0),
-    temp1 REG_v(v1),
-    temp2 REG_v(v2),
-    temp3 REG_v(v3),
-    temp4 REG_v(v4),
-    temp5 REG_v(v5),
-    temp6 REG_v(v6),
-    temp7 REG_v(v7);
-  register vector signed short
-    temp0S REG_v(v8),
-    temp1S REG_v(v9),
-    temp2S REG_v(v10),
-    temp3S REG_v(v11),
-    temp4S REG_v(v12),
-    temp5S REG_v(v13),
-    temp6S REG_v(v14),
-    temp7S REG_v(v15);
-  register const_vector unsigned char vzero REG_v(v31)= (const_vector unsigned char)vec_splat_u8(0);
+    int sum;
+    register vector signed short
+        temp0 REG_v(v0),
+        temp1 REG_v(v1),
+        temp2 REG_v(v2),
+        temp3 REG_v(v3),
+        temp4 REG_v(v4),
+        temp5 REG_v(v5),
+        temp6 REG_v(v6),
+        temp7 REG_v(v7);
+    register vector signed short
+        temp0S REG_v(v8),
+        temp1S REG_v(v9),
+        temp2S REG_v(v10),
+        temp3S REG_v(v11),
+        temp4S REG_v(v12),
+        temp5S REG_v(v13),
+        temp6S REG_v(v14),
+        temp7S REG_v(v15);
+    register const_vector unsigned char vzero REG_v(v31)=
+        (const_vector unsigned char)vec_splat_u8(0);
   {
-    register const_vector signed short vprod1 REG_v(v16)= (const_vector signed short)AVV( 1,-1, 1,-1, 1,-1, 1,-1);
-    register const_vector signed short vprod2 REG_v(v17)= (const_vector signed short)AVV( 1, 1,-1,-1, 1, 1,-1,-1);
-    register const_vector signed short vprod3 REG_v(v18)= (const_vector signed short)AVV( 1, 1, 1, 1,-1,-1,-1,-1);
-    register const_vector unsigned char perm1 REG_v(v19)= (const_vector unsigned char)
-      AVV(0x02, 0x03, 0x00, 0x01,
-       0x06, 0x07, 0x04, 0x05,
-       0x0A, 0x0B, 0x08, 0x09,
-       0x0E, 0x0F, 0x0C, 0x0D);
-    register const_vector unsigned char perm2 REG_v(v20)= (const_vector unsigned char)
-      AVV(0x04, 0x05, 0x06, 0x07,
-       0x00, 0x01, 0x02, 0x03,
-       0x0C, 0x0D, 0x0E, 0x0F,
-       0x08, 0x09, 0x0A, 0x0B);
-    register const_vector unsigned char perm3 REG_v(v21)= (const_vector unsigned char)
-      AVV(0x08, 0x09, 0x0A, 0x0B,
-       0x0C, 0x0D, 0x0E, 0x0F,
-       0x00, 0x01, 0x02, 0x03,
-       0x04, 0x05, 0x06, 0x07);
+    register const_vector signed short vprod1 REG_v(v16)=
+        (const_vector signed short)AVV( 1,-1, 1,-1, 1,-1, 1,-1);
+    register const_vector signed short vprod2 REG_v(v17)=
+        (const_vector signed short)AVV( 1, 1,-1,-1, 1, 1,-1,-1);
+    register const_vector signed short vprod3 REG_v(v18)=
+        (const_vector signed short)AVV( 1, 1, 1, 1,-1,-1,-1,-1);
+    register const_vector unsigned char perm1 REG_v(v19)=
+        (const_vector unsigned char)
+        AVV(0x02, 0x03, 0x00, 0x01, 0x06, 0x07, 0x04, 0x05,
+            0x0A, 0x0B, 0x08, 0x09, 0x0E, 0x0F, 0x0C, 0x0D);
+    register const_vector unsigned char perm2 REG_v(v20)=
+        (const_vector unsigned char)
+        AVV(0x04, 0x05, 0x06, 0x07, 0x00, 0x01, 0x02, 0x03,
+            0x0C, 0x0D, 0x0E, 0x0F, 0x08, 0x09, 0x0A, 0x0B);
+    register const_vector unsigned char perm3 REG_v(v21)=
+        (const_vector unsigned char)
+        AVV(0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
+            0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07);
 
 #define ONEITERBUTTERFLY(i, res1, res2)                                 \
     {                                                                   \
@@ -1642,27 +1642,27 @@
 
 int hadamard8_diff16_altivec(/*MpegEncContext*/ void *s, uint8_t *dst, uint8_t *src, int stride, int h){
 POWERPC_PERF_DECLARE(altivec_hadamard8_diff16_num, 1);
-  int score;
+    int score;
 POWERPC_PERF_START_COUNT(altivec_hadamard8_diff16_num, 1);
-  score = hadamard8_diff16x8_altivec(s, dst, src, stride, 8);
-  if (h==16) {
-    dst += 8*stride;
-    src += 8*stride;
-    score += hadamard8_diff16x8_altivec(s, dst, src, stride, 8);
-  }
+    score = hadamard8_diff16x8_altivec(s, dst, src, stride, 8);
+    if (h==16) {
+        dst += 8*stride;
+        src += 8*stride;
+        score += hadamard8_diff16x8_altivec(s, dst, src, stride, 8);
+    }
 POWERPC_PERF_STOP_COUNT(altivec_hadamard8_diff16_num, 1);
-  return score;
+    return score;
 }
 
 int has_altivec(void)
 {
 #ifdef __AMIGAOS4__
-        ULONG result = 0;
-        extern struct ExecIFace *IExec;
+    ULONG result = 0;
+    extern struct ExecIFace *IExec;
 
-        IExec->GetCPUInfoTags(GCIT_VectorUnit, &result, TAG_DONE);
-        if (result == VECTORTYPE_ALTIVEC) return 1;
-        return 0;
+    IExec->GetCPUInfoTags(GCIT_VectorUnit, &result, TAG_DONE);
+    if (result == VECTORTYPE_ALTIVEC) return 1;
+    return 0;
 #else /* __AMIGAOS4__ */
 
 #ifdef CONFIG_DARWIN
@@ -1757,75 +1757,66 @@
  }
 POWERPC_PERF_STOP_COUNT(altivec_avg_pixels8_xy2_num, 1);
 #else /* ALTIVEC_USE_REFERENCE_C_CODE */
-   register int i;
-   register vector unsigned char
-     pixelsv1, pixelsv2,
-     pixelsavg;
-   register vector unsigned char
-     blockv, temp1, temp2, blocktemp;
-   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);
+    register int i;
+    register vector unsigned char pixelsv1, pixelsv2, pixelsavg;
+    register vector unsigned char blockv, temp1, temp2, blocktemp;
+    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);
-   pixelssum1 = vec_add((vector unsigned short)pixelsv1,
-                        (vector unsigned short)pixelsv2);
-   pixelssum1 = vec_add(pixelssum1, vctwo);
+    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);
+    pixelssum1 = vec_add((vector unsigned short)pixelsv1,
+                         (vector unsigned short)pixelsv2);
+    pixelssum1 = vec_add(pixelssum1, vctwo);
 
 POWERPC_PERF_START_COUNT(altivec_avg_pixels8_xy2_num, 1);
-   for (i = 0; i < h ; i++) {
-     int rightside = ((unsigned long)block & 0x0000000F);
-     blockv = vec_ld(0, block);
+    for (i = 0; i < h ; i++) {
+        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));
-     }
+        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);
-     pixelssum2 = vec_add((vector unsigned short)pixelsv1,
-                          (vector unsigned short)pixelsv2);
-     temp3 = vec_add(pixelssum1, pixelssum2);
-     temp3 = vec_sra(temp3, vctwo);
-     pixelssum1 = vec_add(pixelssum2, vctwo);
-     pixelsavg = vec_packsu(temp3, (vector unsigned short) vczero);
+        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);
+        temp3 = vec_sra(temp3, vctwo);
+        pixelssum1 = vec_add(pixelssum2, vctwo);
+        pixelsavg = vec_packsu(temp3, (vector unsigned short) vczero);
 
-     if (rightside)
-     {
-       blocktemp = vec_perm(blockv, pixelsavg, vcprm(0, 1, s0, s1));
-     }
-     else
-     {
-       blocktemp = vec_perm(blockv, pixelsavg, vcprm(s0, s1, 2, 3));
-     }
+        if (rightside) {
+            blocktemp = vec_perm(blockv, pixelsavg, vcprm(0, 1, s0, s1));
+        } else {
+            blocktemp = vec_perm(blockv, pixelsavg, vcprm(s0, s1, 2, 3));
+        }
 
-     blockv = vec_avg(blocktemp, blockv);
-     vec_st(blockv, 0, block);
+        blockv = vec_avg(blocktemp, blockv);
+        vec_st(blockv, 0, block);
 
-     block += line_size;
-     pixels += line_size;
-   }
+        block += line_size;
+        pixels += line_size;
+    }
 
 POWERPC_PERF_STOP_COUNT(altivec_avg_pixels8_xy2_num, 1);
 #endif /* ALTIVEC_USE_REFERENCE_C_CODE */