diff options
Diffstat (limited to 'src/post/goom/xmmx.c')
| -rw-r--r-- | src/post/goom/xmmx.c | 168 |
1 files changed, 83 insertions, 85 deletions
diff --git a/src/post/goom/xmmx.c b/src/post/goom/xmmx.c index b0afc31a6..ed82172b8 100644 --- a/src/post/goom/xmmx.c +++ b/src/post/goom/xmmx.c @@ -1,12 +1,10 @@ -/* a definir pour avoir exactement le meme resultat que la fonction C - * (un chouillat plus lent) - */ #ifdef HAVE_MMX -#define STRICT_COMPAT - -//#define HAVE_ATHLON +/* a definir pour avoir exactement le meme resultat que la fonction C + * (un chouillat plus lent).. mais la difference est assez peu notable. + */ +// #define STRICT_COMPAT #define BUFFPOINTNB 16 #define BUFFPOINTMASK 0xffff @@ -50,10 +48,12 @@ void zoom_filter_xmmx (int prevX, int prevY, ratiox.d[0] = buffratio; ratiox.d[1] = buffratio; - movq_m2r (ratiox, mm6); - pslld_i2r (16,mm6); - pxor_r2r (mm7,mm7); /* mise a zero de mm7 */ + asm volatile + ("\n\t movq %[ratio], %%mm6" + "\n\t pslld $16, %%mm6" /* mm6 = [rat16=buffratio<<16 | rat16=buffratio<<16] */ + "\n\t pxor %%mm7, %%mm7" /* mm7 = 0 */ + ::[ratio]"m"(ratiox)); loop=0; @@ -62,30 +62,30 @@ void zoom_filter_xmmx (int prevX, int prevY, */ while (loop < bufsize) { - /* - * pre : mm6 = [buffratio<<16|buffratio<<16] - * post : mm0 = S + ((D-S)*buffratio)>>16 format [X|Y] + /* Thread #1 + * pre : mm6 = [rat16|rat16] + * post : mm0 = S + ((D-S)*rat16 format [X|Y] * modified = mm0,mm1,mm2 */ - __asm__ __volatile__ ( - "movq %0,%%mm0\n" - "movq %1,%%mm1\n" - : :"X"(brutS[loop]),"X"(brutD[loop]) - ); /* mm0 = S */ - - psubd_r2r (mm0,mm1); /* mm1 = D - S */ - movq_r2r (mm1, mm2); /* mm2 = D - S */ - - pslld_i2r (16,mm1); - mmx_r2r (pmulhuw, mm6, mm1); /* mm1 = ?? */ - pmullw_r2r (mm6, mm2); - - paddd_r2r (mm2, mm1); /* mm1 = (D - S) * buffratio >> 16 */ - pslld_i2r (16,mm0); - - paddd_r2r (mm1, mm0); /* mm0 = S + mm1 */ - psrld_i2r (16, mm0); + asm volatile + ("#1 \n\t movq %[brutS], %%mm0" + "#1 \n\t movq %[brutD], %%mm1" + "#1 \n\t psubd %%mm0, %%mm1" /* mm1 = D - S */ + "#1 \n\t movq %%mm1, %%mm2" /* mm2 = D - S */ + "#1 \n\t pslld $16, %%mm1" + "#1 \n\t pmullw %%mm6, %%mm2" + "#1 \n\t pmulhuw %%mm6, %%mm1" + + "#1 \n\t pslld $16, %%mm0" + "#1 \n\t paddd %%mm2, %%mm1" /* mm1 = (D - S) * buffratio >> 16 */ + + "#1 \n\t paddd %%mm1, %%mm0" /* mm0 = S + mm1 */ + "#1 \n\t psrld $16, %%mm0" + : + : [brutS]"g"(brutS[loop]) + , [brutD]"g"(brutD[loop]) + ); /* mm0 = S */ /* * pre : mm0 : position vector on screen @@ -93,45 +93,48 @@ void zoom_filter_xmmx (int prevX, int prevY, * post : clipped mm0 * modified : mm0,mm1,mm2 */ - movq_m2r (prevXY,mm1); - pcmpgtd_r2r (mm0, mm1); /* mm0 en X contient : - 1111 si prevXY > px - 0000 si prevXY <= px - (idem pour y) */ + asm volatile + ("#1 \n\t movq %[prevXY], %%mm1" + "#1 \n\t pcmpgtd %%mm0, %%mm1" + /* mm0 en X contient (idem pour Y) : + * 1111 si prevXY > px + * 0000 si prevXY <= px */ #ifdef STRICT_COMPAT - movq_r2r (mm1,mm2); - punpckhdq_r2r (mm2,mm2); - punpckldq_r2r (mm1,mm1); - pand_r2r (mm2, mm0); + "#1 \n\t movq %%mm1, %%mm2" + "#1 \n\t punpckhdq %%mm2, %%mm2" + "#1 \n\t punpckldq %%mm1, %%mm1" + "#1 \n\t pand %%mm2, %%mm0" #endif - pand_r2r (mm1, mm0); /* on met a zero la partie qui deborde */ - /* - * pre : mm0 : clipped position on screen + "#1 \n\t pand %%mm1, %%mm0" /* on met a zero la partie qui deborde */ + ::[prevXY]"m"(prevXY)); + + /* Thread #2 + * pre : mm0 : clipped position on screen * - * post : mm3 & mm4 : coefs for this position - * mm1 : X vector [0|X] + * post : mm3 : coefs for this position + * mm1 : X vector [0|X] * - * modif : eax,ecx + * modif : eax,esi */ __asm__ __volatile__ ( - "movd %%mm0,%%ecx\n" - "movq %%mm0,%%mm1\n" + "#2 \n\t movd %%mm0,%%esi" + "#2 \n\t movq %%mm0,%%mm1" - "andl $15,%%ecx\n" - "psrlq $32,%%mm1\n" + "#2 \n\t andl $15,%%esi" + "#2 \n\t psrlq $32,%%mm1" - "shll $6,%%ecx\n" - "movd %%mm1,%%eax\n" + "#2 \n\t shll $6,%%esi" + "#2 \n\t movd %%mm1,%%eax" - "addl %0,%%ecx\n" - "andl $15,%%eax\n" + "#2 \n\t addl %[precalCoef],%%esi" + "#2 \n\t andl $15,%%eax" - "movd (%%ecx,%%eax,4),%%mm3\n" - ::"g"(precalCoef):"eax","ecx"); + "#2 \n\t movd (%%esi,%%eax,4),%%mm3" + ::[precalCoef]"g"(precalCoef):"eax","esi"); /* - * extraction des coefficients... + * extraction des coefficients... (Thread #3) * * pre : coef dans mm3 * @@ -141,49 +144,44 @@ void zoom_filter_xmmx (int prevX, int prevY, * modif : mm5 */ - /* entrelace avec portion d'apres (cf les '^') - movq_r2r (mm3, mm5); / * ??-??-??-??-c4-c3-c2-c1 * / - punpcklbw_r2r (mm5, mm3); / * c4-c4-c3-c3-c2-c2-c1-c1 * / - movq_r2r (mm3, mm4); / * c4-c4-c3-c3-c2-c2-c1-c1 * / - movq_r2r (mm3, mm5); / * c4-c4-c3-c3-c2-c2-c1-c1 * / - - punpcklbw_r2r (mm5, mm3); / * c2-c2-c2-c2-c1-c1-c1-c1 * / - punpckhbw_r2r (mm5, mm4); / * c4-c4-c4-c4-c3-c3-c3-c3 * / - */ - - /* + /* (Thread #4) * pre : mm0 : Y pos [*|Y] * mm1 : X pos [*|X] * * post : mm0 : expix1[position] * mm2 : expix1[position+largeur] * - * modif : eax,ecx + * modif : eax, esi */ - psrld_i2r (PERTEDEC,mm0); - psrld_i2r (PERTEDEC,mm1); __asm__ __volatile__ ( - "movd %%mm1,%%eax\n" - /*^*/ "movq %%mm3,%%mm5\n" /*^*/ + "#2 \n\t psrld $4, %%mm0" + "#2 \n\t psrld $4, %%mm1" /* PERTEDEC = $4 */ + + "#4 \n\t movd %%mm1,%%eax" + "#3 \n\t movq %%mm3,%%mm5" + + "#4 \n\t mull %[prevX]" + "#4 \n\t movd %%mm0,%%esi" - "mull %1\n" - "movd %%mm0,%%ecx\n" - /*^*/ "punpcklbw %%mm5, %%mm3\n" /*^*/ + "#3 \n\t punpcklbw %%mm5, %%mm3" + "#4 \n\t addl %%esi, %%eax" - "addl %%ecx,%%eax\n" - /*^*/ "movq %%mm3,%%mm4\n" /*^*/ - /*^*/ "movq %%mm3,%%mm5\n" /*^*/ + "#3 \n\t movq %%mm3, %%mm4" + "#3 \n\t movq %%mm3, %%mm5" - "movl %0,%%ecx\n" - /*^*/ "punpcklbw %%mm5,%%mm3\n" /*^*/ + "#4 \n\t movl %[expix1], %%esi" + "#3 \n\t punpcklbw %%mm5, %%mm3" - "movq (%%ecx,%%eax,4),%%mm0\n" - /*^*/ "punpckhbw %%mm5,%%mm4\n" /*^*/ + "#4 \n\t movq (%%esi,%%eax,4),%%mm0" + "#3 \n\t punpckhbw %%mm5, %%mm4" - "addl %1,%%eax\n" - "movq (%%ecx,%%eax,4),%%mm2\n" + "#4 \n\t addl %[prevX],%%eax" + "#4 \n\t movq (%%esi,%%eax,4),%%mm2" - : : "X"(expix1), "X"(prevX):"eax","ecx" + : + : [expix1] "g"(expix1) + , [prevX] "g"(prevX) + :"eax","esi" ); /* |