diff options
author | Romain Dolbeau <dolbeau@irisa.fr> | 2003-01-19 19:00:45 +0000 |
---|---|---|
committer | Michael Niedermayer <michaelni@gmx.at> | 2003-01-19 19:00:45 +0000 |
commit | 35e5fb067e0099fde6c3429131f11433eb92f22b (patch) | |
tree | 50480aa9498173e6b2f12f977d6e948b8f90fa53 /libavcodec/ppc | |
parent | 8c55915b14897fd9bdaae9241f2dd54c00a869ef (diff) | |
download | ffmpeg-35e5fb067e0099fde6c3429131f11433eb92f22b.tar.gz |
PPC perf, PPC clear_block, AltiVec put_pixels8_xy2 patch by (Romain Dolbeau <dolbeau at irisa dot fr>)
Originally committed as revision 1476 to svn://svn.ffmpeg.org/ffmpeg/trunk
Diffstat (limited to 'libavcodec/ppc')
-rw-r--r-- | libavcodec/ppc/dsputil_altivec.c | 254 | ||||
-rw-r--r-- | libavcodec/ppc/dsputil_altivec.h | 67 | ||||
-rw-r--r-- | libavcodec/ppc/dsputil_ppc.c | 167 | ||||
-rw-r--r-- | libavcodec/ppc/dsputil_ppc.h | 120 | ||||
-rw-r--r-- | libavcodec/ppc/fft_altivec.c | 10 | ||||
-rw-r--r-- | libavcodec/ppc/gmc_altivec.c | 10 | ||||
-rw-r--r-- | libavcodec/ppc/idct_altivec.c | 20 | ||||
-rw-r--r-- | libavcodec/ppc/mpegvideo_altivec.c | 6 |
8 files changed, 513 insertions, 141 deletions
diff --git a/libavcodec/ppc/dsputil_altivec.c b/libavcodec/ppc/dsputil_altivec.c index e7f34ee16c..fcae7c1180 100644 --- a/libavcodec/ppc/dsputil_altivec.c +++ b/libavcodec/ppc/dsputil_altivec.c @@ -24,22 +24,6 @@ #include <sys/sysctl.h> #endif -#ifdef ALTIVEC_TBL_PERFORMANCE_REPORT -unsigned long long perfdata[altivec_perf_total][altivec_data_total]; -/* list below must match enum in dsputil_altivec.h */ -static unsigned char* perfname[] = { - "fft_calc", - "gmc1", - "dct_unquantize_h263", - "idct_add", - "idct_put", - "put_pixels_clamped", - "put_pixels16", - "avg_pixels16" -}; -#include <stdio.h> -#endif - int pix_abs16x16_x2_altivec(uint8_t *pix1, uint8_t *pix2, int line_size) { int i; @@ -648,12 +632,12 @@ extern UINT8 cropTbl[]; void put_pixels_clamped_altivec(const DCTELEM *block, UINT8 *restrict pixels, int line_size) { -ALTIVEC_TBL_DECLARE(altivec_put_pixels_clamped_num, 1); +POWERPC_TBL_DECLARE(altivec_put_pixels_clamped_num, 1); #ifdef ALTIVEC_USE_REFERENCE_C_CODE int i; UINT8 *cm = cropTbl + MAX_NEG_CROP; -ALTIVEC_TBL_START_COUNT(altivec_put_pixels_clamped_num, 1); +POWERPC_TBL_START_COUNT(altivec_put_pixels_clamped_num, 1); /* read the pixels */ for(i=0;i<8;i++) { @@ -670,7 +654,7 @@ ALTIVEC_TBL_START_COUNT(altivec_put_pixels_clamped_num, 1); block += 8; } -ALTIVEC_TBL_STOP_COUNT(altivec_put_pixels_clamped_num, 1); +POWERPC_TBL_STOP_COUNT(altivec_put_pixels_clamped_num, 1); #else /* ALTIVEC_USE_REFERENCE_C_CODE */ register const vector short vczero = (const vector short)(0); @@ -681,7 +665,7 @@ ALTIVEC_TBL_STOP_COUNT(altivec_put_pixels_clamped_num, 1); pixelsv0, pixelsv1, pixelsv2, pixelsv3, pixelsv4, pixelsv0old, pixelsv4old; -ALTIVEC_TBL_START_COUNT(altivec_put_pixels_clamped_num, 1); +POWERPC_TBL_START_COUNT(altivec_put_pixels_clamped_num, 1); blockv0 = vec_ld(0, block); blockv1 = vec_ld(16, block); @@ -720,17 +704,17 @@ ALTIVEC_TBL_START_COUNT(altivec_put_pixels_clamped_num, 1); vec_st(pixelsv3, 48, pixels); } -ALTIVEC_TBL_STOP_COUNT(altivec_put_pixels_clamped_num, 1); +POWERPC_TBL_STOP_COUNT(altivec_put_pixels_clamped_num, 1); #endif /* ALTIVEC_USE_REFERENCE_C_CODE */ } void put_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h) { -ALTIVEC_TBL_DECLARE(altivec_put_pixels16_num, 1); +POWERPC_TBL_DECLARE(altivec_put_pixels16_num, 1); #ifdef ALTIVEC_USE_REFERENCE_C_CODE int i; -ALTIVEC_TBL_START_COUNT(altivec_put_pixels16_num, 1); +POWERPC_TBL_START_COUNT(altivec_put_pixels16_num, 1); for(i=0; i<h; i++) { *((uint32_t*)(block )) = (((const struct unaligned_32 *) (pixels))->l); @@ -741,25 +725,24 @@ ALTIVEC_TBL_START_COUNT(altivec_put_pixels16_num, 1); block +=line_size; } -ALTIVEC_TBL_STOP_COUNT(altivec_put_pixels16_num, 1); +POWERPC_TBL_STOP_COUNT(altivec_put_pixels16_num, 1); #else /* ALTIVEC_USE_REFERENCE_C_CODE */ - - register vector unsigned char perm = vec_lvsl(0, pixels); register vector unsigned char pixelsv1, pixelsv2; int i; -ALTIVEC_TBL_START_COUNT(altivec_put_pixels16_num, 1); +POWERPC_TBL_START_COUNT(altivec_put_pixels16_num, 1); for(i=0; i<h; i++) { pixelsv1 = vec_ld(0, (unsigned char*)pixels); pixelsv2 = vec_ld(16, (unsigned char*)pixels); - vec_st(vec_perm(pixelsv1, pixelsv2, perm), 0, (unsigned char*)block); + vec_st(vec_perm(pixelsv1, pixelsv2, vec_lvsl(0, pixels)), + 0, (unsigned char*)block); pixels+=line_size; block +=line_size; } -ALTIVEC_TBL_STOP_COUNT(altivec_put_pixels16_num, 1); +POWERPC_TBL_STOP_COUNT(altivec_put_pixels16_num, 1); #endif /* ALTIVEC_USE_REFERENCE_C_CODE */ } @@ -767,11 +750,11 @@ ALTIVEC_TBL_STOP_COUNT(altivec_put_pixels16_num, 1); #define op_avg(a,b) a = ( ((a)|(b)) - ((((a)^(b))&0xFEFEFEFEUL)>>1) ) void avg_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h) { -ALTIVEC_TBL_DECLARE(altivec_avg_pixels16_num, 1); +POWERPC_TBL_DECLARE(altivec_avg_pixels16_num, 1); #ifdef ALTIVEC_USE_REFERENCE_C_CODE int i; -ALTIVEC_TBL_START_COUNT(altivec_avg_pixels16_num, 1); +POWERPC_TBL_START_COUNT(altivec_avg_pixels16_num, 1); for(i=0; i<h; i++) { op_avg(*((uint32_t*)(block)),(((const struct unaligned_32 *)(pixels))->l)); @@ -782,32 +765,210 @@ ALTIVEC_TBL_START_COUNT(altivec_avg_pixels16_num, 1); block +=line_size; } -ALTIVEC_TBL_STOP_COUNT(altivec_avg_pixels16_num, 1); +POWERPC_TBL_STOP_COUNT(altivec_avg_pixels16_num, 1); #else /* ALTIVEC_USE_REFERENCE_C_CODE */ - - register vector unsigned char perm = vec_lvsl(0, pixels); register vector unsigned char pixelsv1, pixelsv2, pixelsv, blockv; int i; -ALTIVEC_TBL_START_COUNT(altivec_avg_pixels16_num, 1); +POWERPC_TBL_START_COUNT(altivec_avg_pixels16_num, 1); for(i=0; i<h; i++) { pixelsv1 = vec_ld(0, (unsigned char*)pixels); pixelsv2 = vec_ld(16, (unsigned char*)pixels); blockv = vec_ld(0, block); - pixelsv = vec_perm(pixelsv1, pixelsv2, perm); + pixelsv = vec_perm(pixelsv1, pixelsv2, vec_lvsl(0, pixels)); blockv = vec_avg(blockv,pixelsv); vec_st(blockv, 0, (unsigned char*)block); pixels+=line_size; block +=line_size; } -ALTIVEC_TBL_STOP_COUNT(altivec_avg_pixels16_num, 1); +POWERPC_TBL_STOP_COUNT(altivec_avg_pixels16_num, 1); #endif /* ALTIVEC_USE_REFERENCE_C_CODE */ } +void avg_pixels8_altivec(uint8_t * block, const uint8_t * pixels, + int line_size, int h) +{ +POWERPC_TBL_DECLARE(altivec_avg_pixels8_num, 1); +#ifdef ALTIVEC_USE_REFERENCE_C_CODE + int i; +POWERPC_TBL_START_COUNT(altivec_avg_pixels8_num, 1); + for (i = 0; i < h; i++) { + *((uint32_t *) (block)) = + (((*((uint32_t *) (block))) | + ((((const struct unaligned_32 *) (pixels))->l))) - + ((((*((uint32_t *) (block))) ^ + ((((const struct unaligned_32 *) (pixels))-> + l))) & 0xFEFEFEFEUL) >> 1)); + *((uint32_t *) (block + 4)) = + (((*((uint32_t *) (block + 4))) | + ((((const struct unaligned_32 *) (pixels + 4))->l))) - + ((((*((uint32_t *) (block + 4))) ^ + ((((const struct unaligned_32 *) (pixels + + 4))-> + l))) & 0xFEFEFEFEUL) >> 1)); + pixels += line_size; + block += line_size; + } +POWERPC_TBL_STOP_COUNT(altivec_avg_pixels8_num, 1); + +#else /* ALTIVEC_USE_REFERENCE_C_CODE */ + register vector unsigned char pixelsv1, pixelsv2, pixelsv, blockv; + int i; + +POWERPC_TBL_START_COUNT(altivec_avg_pixels8_num, 1); + + for (i = 0; i < h; i++) { + /* + block is 8 bytes-aligned, so we're either in the + left block (16 bytes-aligned) or in the right block (not) + */ + int rightside = ((unsigned long)block & 0x0000000F); + + blockv = vec_ld(0, block); + pixelsv1 = vec_ld(0, (unsigned char*)pixels); + pixelsv2 = vec_ld(16, (unsigned char*)pixels); + pixelsv = vec_perm(pixelsv1, pixelsv2, vec_lvsl(0, pixels)); + + if (rightside) + { + pixelsv = vec_perm(blockv, pixelsv, vcprm(0,1,s0,s1)); + } + else + { + pixelsv = vec_perm(blockv, pixelsv, vcprm(s0,s1,2,3)); + } + + blockv = vec_avg(blockv, pixelsv); + + vec_st(blockv, 0, block); + + pixels += line_size; + block += line_size; + } + +POWERPC_TBL_STOP_COUNT(altivec_avg_pixels8_num, 1); + +#endif /* ALTIVEC_USE_REFERENCE_C_CODE */ +} + +void put_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h) +{ +POWERPC_TBL_DECLARE(altivec_put_pixels8_xy2_num, 1); +#ifdef ALTIVEC_USE_REFERENCE_C_CODE + int j; +POWERPC_TBL_START_COUNT(altivec_put_pixels8_xy2_num, 1); + for (j = 0; j < 2; j++) { + int i; + const uint32_t a = (((const struct unaligned_32 *) (pixels))->l); + const uint32_t b = + (((const struct unaligned_32 *) (pixels + 1))->l); + uint32_t l0 = + (a & 0x03030303UL) + (b & 0x03030303UL) + 0x02020202UL; + uint32_t h0 = + ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2); + uint32_t l1, h1; + pixels += line_size; + for (i = 0; i < h; i += 2) { + uint32_t a = (((const struct unaligned_32 *) (pixels))->l); + uint32_t b = (((const struct unaligned_32 *) (pixels + 1))->l); + l1 = (a & 0x03030303UL) + (b & 0x03030303UL); + h1 = ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2); + *((uint32_t *) block) = + h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL); + pixels += line_size; + block += line_size; + a = (((const struct unaligned_32 *) (pixels))->l); + b = (((const struct unaligned_32 *) (pixels + 1))->l); + l0 = (a & 0x03030303UL) + (b & 0x03030303UL) + 0x02020202UL; + h0 = ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2); + *((uint32_t *) block) = + h0 + h1 + (((l0 + l1) >> 2) & 0x0F0F0F0FUL); + pixels += line_size; + block += line_size; + } pixels += 4 - line_size * (h + 1); + block += 4 - line_size * h; + } + +POWERPC_TBL_STOP_COUNT(altivec_put_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; + register vector unsigned short + pixelssum1, pixelssum2, temp3; + register const vector unsigned char vczero = (const vector unsigned char)(0); + register const vector unsigned short vctwo = (const vector unsigned short)(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); + +POWERPC_TBL_START_COUNT(altivec_put_pixels8_xy2_num, 1); + 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)); + } + + 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) + { + blockv = vec_perm(blockv, pixelsavg, vcprm(0, 1, s0, s1)); + } + else + { + blockv = vec_perm(blockv, pixelsavg, vcprm(s0, s1, 2, 3)); + } + + vec_st(blockv, 0, block); + + block += line_size; + pixels += line_size; + } + +POWERPC_TBL_STOP_COUNT(altivec_put_pixels8_xy2_num, 1); +#endif /* ALTIVEC_USE_REFERENCE_C_CODE */ +} + int has_altivec(void) { #if CONFIG_DARWIN @@ -822,22 +983,3 @@ int has_altivec(void) #endif return 0; } - -#ifdef ALTIVEC_TBL_PERFORMANCE_REPORT -void altivec_display_perf_report(void) -{ - int i; - fprintf(stderr, "AltiVec performance report\n Values are from the Time Base register, and represent 4 bus cycles.\n"); - for(i = 0 ; i < altivec_perf_total ; i++) - { - if (perfdata[i][altivec_data_num] != (unsigned long long)0) - fprintf(stderr, " Function \"%s\":\n\tmin: %llu\n\tmax: %llu\n\tavg: %1.2lf (%llu)\n", - perfname[i], - perfdata[i][altivec_data_min], - perfdata[i][altivec_data_max], - (double)perfdata[i][altivec_data_sum] / - (double)perfdata[i][altivec_data_num], - perfdata[i][altivec_data_num]); - } -} -#endif /* ALTIVEC_TBL_PERFORMANCE_REPORT */ diff --git a/libavcodec/ppc/dsputil_altivec.h b/libavcodec/ppc/dsputil_altivec.h index b7b4238db9..1f66017270 100644 --- a/libavcodec/ppc/dsputil_altivec.h +++ b/libavcodec/ppc/dsputil_altivec.h @@ -17,6 +17,13 @@ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */ +#ifndef _DSPUTIL_ALTIVEC_ +#define _DSPUTIL_ALTIVEC_ + +#include "dsputil_ppc.h" + +#ifdef HAVE_ALTIVEC + extern int pix_abs16x16_x2_altivec(uint8_t *pix1, uint8_t *pix2, int line_size); extern int pix_abs16x16_y2_altivec(uint8_t *pix1, uint8_t *pix2, int line_size); extern int pix_abs16x16_xy2_altivec(uint8_t *pix1, uint8_t *pix2, int line_size); @@ -33,15 +40,15 @@ extern void get_pixels_altivec(DCTELEM* block, const UINT8 * pixels, int line_si extern void add_bytes_altivec(uint8_t *dst, uint8_t *src, int w); extern void put_pixels_clamped_altivec(const DCTELEM *block, UINT8 *restrict pixels, int line_size); -void put_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h); -void avg_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h); +extern void put_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h); +extern void avg_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h); +extern void avg_pixels8_altivec(uint8_t * block, const uint8_t * pixels, int line_size, int h); +extern void put_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h); extern void gmc1_altivec(UINT8 *dst, UINT8 *src, int stride, int h, int x16, int y16, int rounder); extern int has_altivec(void); -#ifdef HAVE_ALTIVEC - // used to build registers permutation vectors (vcprm) // the 's' are for words in the _s_econd vector #define WORD_0 0x00,0x01,0x02,0x03 @@ -67,58 +74,10 @@ extern int has_altivec(void); #define vcii(a,b,c,d) (const vector float)(FLOAT_ ## a, FLOAT_ ## b, FLOAT_ ## c, FLOAT_ ## d) -#ifdef ALTIVEC_TBL_PERFORMANCE_REPORT -void altivec_display_perf_report(void); -/* if you add to the enum below, also add to the perfname array - in dsputil_altivec.c */ -enum altivec_perf_index { - altivec_fft_num = 0, - altivec_gmc1_num, - altivec_dct_unquantize_h263_num, - altivec_idct_add_num, - altivec_idct_put_num, - altivec_put_pixels_clamped_num, - altivec_put_pixels16_num, - altivec_avg_pixels16_num, - altivec_perf_total -}; -enum altivec_data_index { - altivec_data_min = 0, - altivec_data_max, - altivec_data_sum, - altivec_data_num, - altivec_data_total -}; -extern unsigned long long perfdata[altivec_perf_total][altivec_data_total]; -#define ALTIVEC_TBL_DECLARE(a, cond) register unsigned long tbl_start, tbl_stop -#define ALTIVEC_TBL_START_COUNT(a, cond) do { asm("mftb %0" : "=r" (tbl_start)); } while (0) -#define ALTIVEC_TBL_STOP_COUNT(a, cond) do { \ - asm volatile("mftb %0" : "=r" (tbl_stop)); \ - if (tbl_stop > tbl_start) \ - { \ - unsigned long diff = tbl_stop - tbl_start; \ - if (cond) \ - { \ - if (diff < perfdata[a][altivec_data_min]) \ - perfdata[a][altivec_data_min] = diff; \ - if (diff > perfdata[a][altivec_data_max]) \ - perfdata[a][altivec_data_max] = diff; \ - perfdata[a][altivec_data_sum] += diff; \ - perfdata[a][altivec_data_num] ++; \ - } \ - } \ -} while (0) -#else /* ALTIVEC_TBL_PERFORMANCE_REPORT */ -#define ALTIVEC_TBL_DECLARE(a, cond) -#define ALTIVEC_TBL_START_COUNT(a, cond) -#define ALTIVEC_TBL_STOP_COUNT(a, cond) -#endif /* ALTIVEC_TBL_PERFORMANCE_REPORT */ - #else /* HAVE_ALTIVEC */ #ifdef ALTIVEC_USE_REFERENCE_C_CODE #error "I can't use ALTIVEC_USE_REFERENCE_C_CODE if I don't use HAVE_ALTIVEC" #endif /* ALTIVEC_USE_REFERENCE_C_CODE */ -#ifdef ALTIVEC_TBL_PERFORMANCE_REPORT -#error "I can't use ALTIVEC_TBL_PERFORMANCE_REPORT if I don't use HAVE_ALTIVEC" -#endif /* ALTIVEC_TBL_PERFORMANCE_REPORT */ #endif /* HAVE_ALTIVEC */ + +#endif /* _DSPUTIL_ALTIVEC_ */ diff --git a/libavcodec/ppc/dsputil_ppc.c b/libavcodec/ppc/dsputil_ppc.c index ec2dda4690..7dcca85a7c 100644 --- a/libavcodec/ppc/dsputil_ppc.c +++ b/libavcodec/ppc/dsputil_ppc.c @@ -19,6 +19,8 @@ #include "../dsputil.h" +#include "dsputil_ppc.h" + #ifdef HAVE_ALTIVEC #include "dsputil_altivec.h" #endif @@ -36,12 +38,149 @@ int mm_support(void) return result; } +#ifdef POWERPC_TBL_PERFORMANCE_REPORT +unsigned long long perfdata[powerpc_perf_total][powerpc_data_total]; +/* list below must match enum in dsputil_altivec.h */ +static unsigned char* perfname[] = { + "fft_calc_altivec", + "gmc1_altivec", + "dct_unquantize_h263_altivec", + "idct_add_altivec", + "idct_put_altivec", + "put_pixels_clamped_altivec", + "put_pixels16_altivec", + "avg_pixels16_altivec", + "avg_pixels8_altivec", + "put_pixels8_xy2_altivec", + "clear_blocks_dcbz32_ppc" +}; +#ifdef POWERPC_PERF_USE_PMC +unsigned long long perfdata_miss[powerpc_perf_total][powerpc_data_total]; +#endif +#include <stdio.h> +#endif + +#ifdef POWERPC_TBL_PERFORMANCE_REPORT +void powerpc_display_perf_report(void) +{ + int i; +#ifndef POWERPC_PERF_USE_PMC + fprintf(stderr, "AltiVec performance report\n Values are from the Time Base register, and represent 4 bus cycles.\n"); +#else /* POWERPC_PERF_USE_PMC */ + fprintf(stderr, "AltiVec performance report\n Values are from the PMC registers, and represent whatever the registers are set to record.\n"); +#endif /* POWERPC_PERF_USE_PMC */ + for(i = 0 ; i < powerpc_perf_total ; i++) + { + if (perfdata[i][powerpc_data_num] != (unsigned long long)0) + fprintf(stderr, " Function \"%s\" (pmc1):\n\tmin: %llu\n\tmax: %llu\n\tavg: %1.2lf (%llu)\n", + perfname[i], + perfdata[i][powerpc_data_min], + perfdata[i][powerpc_data_max], + (double)perfdata[i][powerpc_data_sum] / + (double)perfdata[i][powerpc_data_num], + perfdata[i][powerpc_data_num]); +#ifdef POWERPC_PERF_USE_PMC + if (perfdata_miss[i][powerpc_data_num] != (unsigned long long)0) + fprintf(stderr, " Function \"%s\" (pmc2):\n\tmin: %llu\n\tmax: %llu\n\tavg: %1.2lf (%llu)\n", + perfname[i], + perfdata_miss[i][powerpc_data_min], + perfdata_miss[i][powerpc_data_max], + (double)perfdata_miss[i][powerpc_data_sum] / + (double)perfdata_miss[i][powerpc_data_num], + perfdata_miss[i][powerpc_data_num]); +#endif + } +} +#endif /* POWERPC_TBL_PERFORMANCE_REPORT */ + +/* ***** WARNING ***** WARNING ***** WARNING ***** */ +/* + clear_blocks_dcbz32_ppc will not work properly + on PowerPC processors with a cache line size + not equal to 32 bytes. + Fortunately all processor used by Apple up to + at least the 7450 (aka second generation G4) + use 32 bytes cache line. + This is due to the use of the 'dcbz' instruction. + It simply clear to zero a single cache line, + so you need to know the cache line size to use it ! + It's absurd, but it's fast... +*/ +void clear_blocks_dcbz32_ppc(DCTELEM *blocks) +{ +POWERPC_TBL_DECLARE(powerpc_clear_blocks_dcbz32, 1); + register int misal = ((unsigned long)blocks & 0x00000010); + register int i = 0; +POWERPC_TBL_START_COUNT(powerpc_clear_blocks_dcbz32, 1); +#if 1 + if (misal) { + ((unsigned long*)blocks)[0] = 0L; + ((unsigned long*)blocks)[1] = 0L; + ((unsigned long*)blocks)[2] = 0L; + ((unsigned long*)blocks)[3] = 0L; + vec_st((vector short)(0), 0, blocks); + i += 16; + } + for ( ; i < sizeof(DCTELEM)*6*64 ; i += 32) { + asm volatile("dcbz %0,%1" : : "r" (blocks), "r" (i) : "memory"); + } + if (misal) { + ((unsigned long*)blocks)[188] = 0L; + ((unsigned long*)blocks)[189] = 0L; + ((unsigned long*)blocks)[190] = 0L; + ((unsigned long*)blocks)[191] = 0L; + i += 16; + } +#else + memset(blocks, 0, sizeof(DCTELEM)*6*64); +#endif +POWERPC_TBL_STOP_COUNT(powerpc_clear_blocks_dcbz32, 1); +} + +/* check dcbz report how many bytes are set to 0 by dcbz */ +long check_dcbz_effect(void) +{ + register char *fakedata = (char*)malloc(1024); + register char *fakedata_middle; + register long zero = 0; + register long i = 0; + long count = 0; + + if (fakedata == NULL) + { + return 0L; + } + + + fakedata_middle = (fakedata + 512); + + memset(fakedata, 0xFF, 1024); + + asm volatile("dcbz %0, %1" : : "r" (fakedata_middle), "r" (zero)); + + for (i = 0; i < 1024 ; i ++) + { + if (fakedata[i] == (char)0) + count++; + } + + free(fakedata); + + return count; +} + void dsputil_init_ppc(DSPContext* c, unsigned mask) { // Common optimisations whether Altivec or not - // ... pending ... - + switch (check_dcbz_effect()) { + case 32: + c->clear_blocks = clear_blocks_dcbz32_ppc; + break; + default: + break; + } + #if HAVE_ALTIVEC if (has_altivec()) { mm_flags |= MM_ALTIVEC; @@ -67,17 +206,29 @@ void dsputil_init_ppc(DSPContext* c, unsigned mask) #endif c->put_pixels_tab[0][0] = put_pixels16_altivec; c->avg_pixels_tab[0][0] = avg_pixels16_altivec; +// next one disabled as it's untested. +#if 0 + c->avg_pixels_tab[1][0] = avg_pixels8_altivec; +#endif + c->put_pixels_tab[1][3] = put_pixels8_xy2_altivec; + c->gmc1 = gmc1_altivec; -#ifdef ALTIVEC_TBL_PERFORMANCE_REPORT +#ifdef POWERPC_TBL_PERFORMANCE_REPORT { int i; - for (i = 0 ; i < altivec_perf_total ; i++) + for (i = 0 ; i < powerpc_perf_total ; i++) { - perfdata[i][altivec_data_min] = 0xFFFFFFFFFFFFFFFF; - perfdata[i][altivec_data_max] = 0x0000000000000000; - perfdata[i][altivec_data_sum] = 0x0000000000000000; - perfdata[i][altivec_data_num] = 0x0000000000000000; + perfdata[i][powerpc_data_min] = 0xFFFFFFFFFFFFFFFF; + perfdata[i][powerpc_data_max] = 0x0000000000000000; + perfdata[i][powerpc_data_sum] = 0x0000000000000000; + perfdata[i][powerpc_data_num] = 0x0000000000000000; +#ifdef POWERPC_PERF_USE_PMC + perfdata_miss[i][powerpc_data_min] = 0xFFFFFFFFFFFFFFFF; + perfdata_miss[i][powerpc_data_max] = 0x0000000000000000; + perfdata_miss[i][powerpc_data_sum] = 0x0000000000000000; + perfdata_miss[i][powerpc_data_num] = 0x0000000000000000; +#endif } } #endif diff --git a/libavcodec/ppc/dsputil_ppc.h b/libavcodec/ppc/dsputil_ppc.h new file mode 100644 index 0000000000..e0611ff886 --- /dev/null +++ b/libavcodec/ppc/dsputil_ppc.h @@ -0,0 +1,120 @@ +/* + * Copyright (c) 2003 Romain Dolbeau <romain@dolbeau.org> + * + * This library 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 of the License, or (at your option) any later version. + * + * This library 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 this library; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + */ + +#ifndef _DSPUTIL_PPC_ +#define _DSPUTIL_PPC_ + +#ifdef POWERPC_TBL_PERFORMANCE_REPORT +void powerpc_display_perf_report(void); +/* if you add to the enum below, also add to the perfname array + in dsputil_ppc.c */ +enum powerpc_perf_index { + altivec_fft_num = 0, + altivec_gmc1_num, + altivec_dct_unquantize_h263_num, + altivec_idct_add_num, + altivec_idct_put_num, + altivec_put_pixels_clamped_num, + altivec_put_pixels16_num, + altivec_avg_pixels16_num, + altivec_avg_pixels8_num, + altivec_put_pixels8_xy2_num, + powerpc_clear_blocks_dcbz32, + powerpc_perf_total +}; +enum powerpc_data_index { + powerpc_data_min = 0, + powerpc_data_max, + powerpc_data_sum, + powerpc_data_num, + powerpc_data_total +}; +extern unsigned long long perfdata[powerpc_perf_total][powerpc_data_total]; +#ifdef POWERPC_PERF_USE_PMC +extern unsigned long long perfdata_miss[powerpc_perf_total][powerpc_data_total]; +#endif + +#ifndef POWERPC_PERF_USE_PMC +#define POWERPC_GET_CYCLES(a) asm volatile("mftb %0" : "=r" (a)) +#define POWERPC_TBL_DECLARE(a, cond) register unsigned long tbl_start, tbl_stop +#define POWERPC_TBL_START_COUNT(a, cond) do { POWERPC_GET_CYCLES(tbl_start); } while (0) +#define POWERPC_TBL_STOP_COUNT(a, cond) do { \ + POWERPC_GET_CYCLES(tbl_stop); \ + if (tbl_stop > tbl_start) \ + { \ + unsigned long diff = tbl_stop - tbl_start; \ + if (cond) \ + { \ + if (diff < perfdata[a][powerpc_data_min]) \ + perfdata[a][powerpc_data_min] = diff; \ + if (diff > perfdata[a][powerpc_data_max]) \ + perfdata[a][powerpc_data_max] = diff; \ + perfdata[a][powerpc_data_sum] += diff; \ + perfdata[a][powerpc_data_num] ++; \ + } \ + } \ +} while (0) + +#else /* POWERPC_PERF_USE_PMC */ +#define POWERPC_GET_CYCLES(a) asm volatile("mfspr %0, 937" : "=r" (a)) +#define POWERPC_GET_MISS(a) asm volatile("mfspr %0, 938" : "=r" (a)) +#define POWERPC_TBL_DECLARE(a, cond) register unsigned long cycles_start, cycles_stop, miss_start, miss_stop; +#define POWERPC_TBL_START_COUNT(a, cond) do { POWERPC_GET_MISS(miss_start); POWERPC_GET_CYCLES(cycles_start); } while (0) +#define POWERPC_TBL_STOP_COUNT(a, cond) do { \ + POWERPC_GET_CYCLES(cycles_stop); \ + POWERPC_GET_MISS(miss_stop); \ + if (cycles_stop >= cycles_start) \ + { \ + unsigned long diff = \ + cycles_stop - cycles_start; \ + if (cond) \ + { \ + if (diff < perfdata[a][powerpc_data_min]) \ + perfdata[a][powerpc_data_min] = diff; \ + if (diff > perfdata[a][powerpc_data_max]) \ + perfdata[a][powerpc_data_max] = diff; \ + perfdata[a][powerpc_data_sum] += diff; \ + perfdata[a][powerpc_data_num] ++; \ + } \ + } \ + if (miss_stop >= miss_start) \ + { \ + unsigned long diff = \ + miss_stop - miss_start; \ + if (cond) \ + { \ + if (diff < perfdata_miss[a][powerpc_data_min]) \ + perfdata_miss[a][powerpc_data_min] = diff; \ + if (diff > perfdata_miss[a][powerpc_data_max]) \ + perfdata_miss[a][powerpc_data_max] = diff; \ + perfdata_miss[a][powerpc_data_sum] += diff; \ + perfdata_miss[a][powerpc_data_num] ++; \ + } \ + } \ +} while (0) + +#endif /* POWERPC_PERF_USE_PMC */ + + +#else /* POWERPC_TBL_PERFORMANCE_REPORT */ +#define POWERPC_TBL_DECLARE(a, cond) +#define POWERPC_TBL_START_COUNT(a, cond) +#define POWERPC_TBL_STOP_COUNT(a, cond) +#endif /* POWERPC_TBL_PERFORMANCE_REPORT */ + +#endif /* _DSPUTIL_PPC_ */ diff --git a/libavcodec/ppc/fft_altivec.c b/libavcodec/ppc/fft_altivec.c index 55bfcdf4d2..db32acd23a 100644 --- a/libavcodec/ppc/fft_altivec.c +++ b/libavcodec/ppc/fft_altivec.c @@ -60,7 +60,7 @@ */ void fft_calc_altivec(FFTContext *s, FFTComplex *z) { -ALTIVEC_TBL_DECLARE(altivec_fft_num, s->nbits >= 6); +POWERPC_TBL_DECLARE(altivec_fft_num, s->nbits >= 6); #ifdef ALTIVEC_USE_REFERENCE_C_CODE int ln = s->nbits; int j, np, np2; @@ -70,7 +70,7 @@ ALTIVEC_TBL_DECLARE(altivec_fft_num, s->nbits >= 6); int l; FFTSample tmp_re, tmp_im; -ALTIVEC_TBL_START_COUNT(altivec_fft_num, s->nbits >= 6); +POWERPC_TBL_START_COUNT(altivec_fft_num, s->nbits >= 6); np = 1 << ln; @@ -135,7 +135,7 @@ ALTIVEC_TBL_START_COUNT(altivec_fft_num, s->nbits >= 6); nloops = nloops << 1; } while (nblocks != 0); -ALTIVEC_TBL_STOP_COUNT(altivec_fft_num, s->nbits >= 6); +POWERPC_TBL_STOP_COUNT(altivec_fft_num, s->nbits >= 6); #else /* ALTIVEC_USE_REFERENCE_C_CODE */ register const vector float vczero = (const vector float)(0.); @@ -147,7 +147,7 @@ ALTIVEC_TBL_STOP_COUNT(altivec_fft_num, s->nbits >= 6); FFTComplex *cptr, *cptr1; int k; -ALTIVEC_TBL_START_COUNT(altivec_fft_num, s->nbits >= 6); +POWERPC_TBL_START_COUNT(altivec_fft_num, s->nbits >= 6); np = 1 << ln; @@ -235,7 +235,7 @@ ALTIVEC_TBL_START_COUNT(altivec_fft_num, s->nbits >= 6); nloops = nloops << 1; } while (nblocks != 0); -ALTIVEC_TBL_STOP_COUNT(altivec_fft_num, s->nbits >= 6); +POWERPC_TBL_STOP_COUNT(altivec_fft_num, s->nbits >= 6); #endif /* ALTIVEC_USE_REFERENCE_C_CODE */ } diff --git a/libavcodec/ppc/gmc_altivec.c b/libavcodec/ppc/gmc_altivec.c index c2d908d039..810a237fe7 100644 --- a/libavcodec/ppc/gmc_altivec.c +++ b/libavcodec/ppc/gmc_altivec.c @@ -28,7 +28,7 @@ */ void gmc1_altivec(UINT8 *dst /* align 8 */, UINT8 *src /* align1 */, int stride, int h, int x16, int y16, int rounder) { -ALTIVEC_TBL_DECLARE(altivec_gmc1_num, h == 8); +POWERPC_TBL_DECLARE(altivec_gmc1_num, h == 8); #ifdef ALTIVEC_USE_REFERENCE_C_CODE const int A=(16-x16)*(16-y16); const int B=( x16)*(16-y16); @@ -36,7 +36,7 @@ ALTIVEC_TBL_DECLARE(altivec_gmc1_num, h == 8); const int D=( x16)*( y16); int i; -ALTIVEC_TBL_START_COUNT(altivec_gmc1_num, h == 8); +POWERPC_TBL_START_COUNT(altivec_gmc1_num, h == 8); for(i=0; i<h; i++) { @@ -52,7 +52,7 @@ ALTIVEC_TBL_START_COUNT(altivec_gmc1_num, h == 8); src+= stride; } -ALTIVEC_TBL_STOP_COUNT(altivec_gmc1_num, h == 8); +POWERPC_TBL_STOP_COUNT(altivec_gmc1_num, h == 8); #else /* ALTIVEC_USE_REFERENCE_C_CODE */ const unsigned short __attribute__ ((aligned(16))) rounder_a[8] = @@ -75,7 +75,7 @@ ALTIVEC_TBL_STOP_COUNT(altivec_gmc1_num, h == 8); unsigned long src_really_odd = (unsigned long)src & 0x0000000F; -ALTIVEC_TBL_START_COUNT(altivec_gmc1_num, h == 8); +POWERPC_TBL_START_COUNT(altivec_gmc1_num, h == 8); tempA = vec_ld(0, (unsigned short*)ABCD); Av = vec_splat(tempA, 0); @@ -163,7 +163,7 @@ ALTIVEC_TBL_START_COUNT(altivec_gmc1_num, h == 8); src += stride; } -ALTIVEC_TBL_STOP_COUNT(altivec_gmc1_num, h == 8); +POWERPC_TBL_STOP_COUNT(altivec_gmc1_num, h == 8); #endif /* ALTIVEC_USE_REFERENCE_C_CODE */ } diff --git a/libavcodec/ppc/idct_altivec.c b/libavcodec/ppc/idct_altivec.c index 533ab3e4c8..41db01f832 100644 --- a/libavcodec/ppc/idct_altivec.c +++ b/libavcodec/ppc/idct_altivec.c @@ -161,16 +161,16 @@ static const vector_s16_t constants[5] = { void idct_put_altivec(uint8_t* dest, int stride, vector_s16_t* block) { -ALTIVEC_TBL_DECLARE(altivec_idct_put_num, 1); +POWERPC_TBL_DECLARE(altivec_idct_put_num, 1); #ifdef ALTIVEC_USE_REFERENCE_C_CODE -ALTIVEC_TBL_START_COUNT(altivec_idct_put_num, 1); +POWERPC_TBL_START_COUNT(altivec_idct_put_num, 1); void simple_idct_put(UINT8 *dest, int line_size, INT16 *block); simple_idct_put(dest, stride, (INT16*)block); -ALTIVEC_TBL_STOP_COUNT(altivec_idct_put_num, 1); +POWERPC_TBL_STOP_COUNT(altivec_idct_put_num, 1); #else /* ALTIVEC_USE_REFERENCE_C_CODE */ vector_u8_t tmp; -ALTIVEC_TBL_START_COUNT(altivec_idct_put_num, 1); +POWERPC_TBL_START_COUNT(altivec_idct_put_num, 1); IDCT @@ -188,18 +188,18 @@ ALTIVEC_TBL_START_COUNT(altivec_idct_put_num, 1); COPY (dest, vx6) dest += stride; COPY (dest, vx7) -ALTIVEC_TBL_STOP_COUNT(altivec_idct_put_num, 1); +POWERPC_TBL_STOP_COUNT(altivec_idct_put_num, 1); #endif /* ALTIVEC_USE_REFERENCE_C_CODE */ } void idct_add_altivec(uint8_t* dest, int stride, vector_s16_t* block) { -ALTIVEC_TBL_DECLARE(altivec_idct_add_num, 1); +POWERPC_TBL_DECLARE(altivec_idct_add_num, 1); #ifdef ALTIVEC_USE_REFERENCE_C_CODE -ALTIVEC_TBL_START_COUNT(altivec_idct_add_num, 1); +POWERPC_TBL_START_COUNT(altivec_idct_add_num, 1); void simple_idct_add(UINT8 *dest, int line_size, INT16 *block); simple_idct_add(dest, stride, (INT16*)block); -ALTIVEC_TBL_STOP_COUNT(altivec_idct_add_num, 1); +POWERPC_TBL_STOP_COUNT(altivec_idct_add_num, 1); #else /* ALTIVEC_USE_REFERENCE_C_CODE */ vector_u8_t tmp; vector_s16_t tmp2, tmp3; @@ -207,7 +207,7 @@ ALTIVEC_TBL_STOP_COUNT(altivec_idct_add_num, 1); vector_u8_t perm1; vector_u8_t p0, p1, p; -ALTIVEC_TBL_START_COUNT(altivec_idct_add_num, 1); +POWERPC_TBL_START_COUNT(altivec_idct_add_num, 1); IDCT @@ -235,7 +235,7 @@ ALTIVEC_TBL_START_COUNT(altivec_idct_add_num, 1); ADD (dest, vx6, perm0) dest += stride; ADD (dest, vx7, perm1) -ALTIVEC_TBL_STOP_COUNT(altivec_idct_add_num, 1); +POWERPC_TBL_STOP_COUNT(altivec_idct_add_num, 1); #endif /* ALTIVEC_USE_REFERENCE_C_CODE */ } diff --git a/libavcodec/ppc/mpegvideo_altivec.c b/libavcodec/ppc/mpegvideo_altivec.c index dfa4c0229a..268aef334f 100644 --- a/libavcodec/ppc/mpegvideo_altivec.c +++ b/libavcodec/ppc/mpegvideo_altivec.c @@ -511,13 +511,13 @@ int dct_quantize_altivec(MpegEncContext* s, void dct_unquantize_h263_altivec(MpegEncContext *s, DCTELEM *block, int n, int qscale) { -ALTIVEC_TBL_DECLARE(altivec_dct_unquantize_h263_num, 1); +POWERPC_TBL_DECLARE(altivec_dct_unquantize_h263_num, 1); int i, level, qmul, qadd; int nCoeffs; assert(s->block_last_index[n]>=0); -ALTIVEC_TBL_START_COUNT(altivec_dct_unquantize_h263_num, 1); +POWERPC_TBL_START_COUNT(altivec_dct_unquantize_h263_num, 1); qadd = (qscale - 1) | 1; qmul = qscale << 1; @@ -630,5 +630,5 @@ ALTIVEC_TBL_START_COUNT(altivec_dct_unquantize_h263_num, 1); } #endif /* ALTIVEC_USE_REFERENCE_C_CODE */ -ALTIVEC_TBL_STOP_COUNT(altivec_dct_unquantize_h263_num, nCoeffs == 63); +POWERPC_TBL_STOP_COUNT(altivec_dct_unquantize_h263_num, nCoeffs == 63); } |