}
now_transferring = true;
__DECL_ALIGNED(32) uint16_t sbuf[16][16];
- __DECL_ALIGNED(32) uint32_t lbuf[16];
- __DECL_ALIGNED(32) uint32_t mbuf[16];
+ __DECL_ALIGNED(32) pair16_t lbuf[16];
+ __DECL_ALIGNED(32) pair16_t mbuf[16];
__DECL_ALIGNED(16) uint16_t pixel_h[8];
__DECL_ALIGNED(16) uint16_t pixel_l[8];
__DECL_ALIGNED(16) uint16_t color_table[16] = {0};
}
}
+ uint32_t vpaddr = (((x - xoffset) % 256 + ((y - yoffset) * 256)) << 1) & 0x1ffff;
if(!(is_halfx) && !(is_halfy)) { // not halfed
- for(int yy = 0; yy < 16; yy++) {
+ int __xstart;
+ int __xend;
+ if(x < xoffset) {
+ __xstart = ((xoffset - x + 16) >= 0) ? (xoffset - x + 16) : 0;
+ __xend = ((xoffset - x + 16) >= 0) ? 16 : 0;
+ } else if(x > (xoffset + 256)) {
+ __xstart = x - (xoffset + 256);
+ if(__xstart >= 16) {
+ __xstart = 0;
+ __xend = 0;
+ } else {
+ __xstart = 0;
+ __xend = 16 - __xstart;
+ }
+ } else { // INSIDE OF WINDOW
+ __xstart = 0;
+ __xend = 16;
+ }
+ if(((__xstart <= 0) && (__xend <= 0))) goto __noop;
+ int __ystart;
+ int __yend;
+ if(y < yoffset) {
+ __ystart = ((yoffset - y + 16) >= 0) ? (yoffset - y + 16) : 0;
+ __yend = ((yoffset - y + 16) >= 0) ? 16 : 0;
+ } else if(y > (yoffset + 256)) {
+ __ystart = y - (yoffset + 256);
+ if(__ystart >= 16) {
+ __ystart = 0;
+ __yend = 0;
+ } else {
+ __ystart = 0;
+ __yend = 16 - __ystart;
+ }
+ } else { // INSIDE OF WINDOW
+ __ystart = 0;
+ __yend = 16;
+ }
+ if(((__ystart <= 0) && (__yend <= 0))) goto __noop;
+ for(int yy = __ystart; yy < __yend; yy++) {
+__DECL_VECTORIZED_LOOP
+ for(int xx = 0; xx < 16; xx++) {
+ lbuf[xx].w = 0;
+ mbuf[xx].w = 0;
+ }
__DECL_VECTORIZED_LOOP
for(int xx = 0; xx < 16; xx++) {
- lbuf[xx] = 0;
- mbuf[xx] = 0;
+ lbuf[xx].w = sbuf[yy][xx];
}
__DECL_VECTORIZED_LOOP
for(int xx = 0; xx < 16; xx++) {
- lbuf[xx] = sbuf[yy][xx];
+ mbuf[xx].w = (lbuf[xx].w == 0) ? 0xffff : 0x0000;
}
- // void __FASTCALL VRAM::write_sprite_data(int x, int y, int xoffset, int yoffset, uint16_t *ptr __assume_aligned(16), int width);
if(d_vram != NULL) {
- //d_vram->write_sprite_data(x, y + yy, xoffset, yoffset, lbuf, 16);
+ uint8_t* vp = d_vram->get_vram_address(vpaddr);
+ __DECL_ALIGNED(16) uint8_t source[32] = {0};
+ if(vp != NULL) {
+ for(int xx = (__xstart << 1); xx < (__xend << 1); xx++) {
+ source[xx] = vp[xx];
+ }
+__DECL_VECTORIZED_LOOP
+ for(int xx = 0; xx < 16; xx++) {
+ source[(xx << 1) + 0] &= mbuf[xx].b.l;
+ source[(xx << 1) + 1] &= mbuf[xx].b.h;
+ }
+__DECL_VECTORIZED_LOOP
+ for(int xx = 0; xx < 16; xx++) {
+ source[(xx << 1) + 0] |= lbuf[xx].b.l;
+ source[(xx << 1) + 1] |= lbuf[xx].b.h;
+ }
+ for(int xx = (__xstart << 1); xx < (__xend << 1) ;xx++) {
+ vp[xx] = source[xx];
+ }
+ }
}
+ vpaddr = (vpaddr + (256 << 1)) & 0x1ffff;
}
- } else if(is_halfx) { // halfx only
- for(int yy = 0; yy < 16; yy++) {
+ } else if((is_halfx) && !(is_halfy)) { // halfx only
+ int __xstart;
+ int __xend;
+ if(x < xoffset) {
+ __xstart = ((xoffset - x + 8) >= 0) ? (xoffset - x + 8) : 0;
+ __xend = ((xoffset - x + 8) >= 0) ? 8 : 0;
+ } else if(x > (xoffset + 256)) {
+ __xstart = x - (xoffset + 256);
+ if(__xstart >= 8) {
+ __xstart = 0;
+ __xend = 0;
+ } else {
+ __xstart = 0;
+ __xend = 8 - __xstart;
+ }
+ } else { // INSIDE OF WINDOW
+ __xstart = 0;
+ __xend = 8;
+ }
+ if(((__xstart <= 0) && (__xend <= 0))) goto __noop;
+ int __ystart;
+ int __yend;
+ if(y < yoffset) {
+ __ystart = ((yoffset - y + 16) >= 0) ? (yoffset - y + 16) : 0;
+ __yend = ((yoffset - y + 16) >= 0) ? 16 : 0;
+ } else if(y > (yoffset + 256)) {
+ __ystart = y - (yoffset + 256);
+ if(__ystart >= 16) {
+ __ystart = 0;
+ __yend = 0;
+ } else {
+ __ystart = 0;
+ __yend = 16 - __ystart;
+ }
+ } else { // INSIDE OF WINDOW
+ __ystart = 0;
+ __yend = 16;
+ }
+ if(((__ystart <= 0) && (__yend <= 0))) goto __noop;
+ for(int yy = __ystart; yy < __yend; yy++) {
__DECL_VECTORIZED_LOOP
for(int xx = 0; xx < 16; xx++) {
- lbuf[xx] = 0;
- mbuf[xx] = 0;
+ lbuf[xx].w = 0;
+ mbuf[xx].w = 0;
}
__DECL_VECTORIZED_LOOP
- for(int xx = 0; xx < 16; xx += 2) {
- lbuf[xx >> 1] += (sbuf[yy][xx] & 0x7fff);
- mbuf[xx >> 1] |= (sbuf[yy][xx] & 0x8000);
+ for(int xx = 0; xx < 16; xx++) {
+ lbuf[xx >> 1].w += (sbuf[yy][xx] & 0x7fff);
+ mbuf[xx >> 1].w |= (sbuf[yy][xx] & 0x8000);
}
__DECL_VECTORIZED_LOOP
for(int xx = 0; xx < 8; xx++) {
- lbuf[xx] = ((lbuf[xx] >> 2) & 0x7fff) | mbuf[xx];
+ lbuf[xx].w = ((lbuf[xx].w >> 1) & 0x7fff) | mbuf[xx].w;
+ }
+__DECL_VECTORIZED_LOOP
+ for(int xx = 0; xx < 8; xx += 1) {
+ mbuf[xx].w = (lbuf[xx].w == 0) ? 0xffff : 0x0000;
}
if(d_vram != NULL) {
- //d_vram->write_sprite_data(x, y + yy, xoffset, yoffset, lbuf, 8);
+ uint8_t* vp = d_vram->get_vram_address(vpaddr);
+ __DECL_ALIGNED(16) uint8_t source[32] = {0};
+ if(vp != NULL) {
+ for(int xx = (__xstart << 1); xx < (__xend << 1); xx++) {
+ source[xx] = vp[xx];
+ }
+__DECL_VECTORIZED_LOOP
+ for(int xx = 0; xx < 8; xx++) {
+ source[(xx << 1) + 0] &= mbuf[xx].b.l;
+ source[(xx << 1) + 1] &= mbuf[xx].b.h;
+ }
+__DECL_VECTORIZED_LOOP
+ for(int xx = 0; xx < 8; xx++) {
+ source[(xx << 1) + 0] |= lbuf[xx].b.l;
+ source[(xx << 1) + 1] |= lbuf[xx].b.h;
+ }
+ for(int xx = (__xstart << 1); xx < (__xend << 1); xx++) {
+ vp[xx] = source[xx];
+ }
+ }
}
+ vpaddr = (vpaddr + (256 << 1)) & 0x1ffff;
}
} else if(is_halfy) { // halfy only
- for(int yy = 0; yy < 16; yy += 2) {
+ int __xstart;
+ int __xend;
+ if(x < xoffset) {
+ __xstart = ((xoffset - x + 16) >= 0) ? (xoffset - x + 16) : 0;
+ __xend = ((xoffset - x + 16) >= 0) ? 16 : 0;
+ } else if(x > (xoffset + 256)) {
+ __xstart = x - (xoffset + 256);
+ if(__xstart >= 16) {
+ __xstart = 0;
+ __xend = 0;
+ } else {
+ __xstart = 0;
+ __xend = 16 - __xstart;
+ }
+ } else { // INSIDE OF WINDOW
+ __xstart = 0;
+ __xend = 16;
+ }
+ if(((__xstart <= 0) && (__xend <= 0))) goto __noop;
+ int __ystart;
+ int __yend;
+ if(y < yoffset) {
+ __ystart = ((yoffset - y + 8) >= 0) ? (yoffset - y + 8) : 0;
+ __yend = ((yoffset - y + 8) >= 0) ? 8 : 0;
+ } else if(y > (yoffset + 256)) {
+ __ystart = y - (yoffset + 256);
+ if(__ystart >= 8) {
+ __ystart = 0;
+ __yend = 0;
+ } else {
+ __ystart = 0;
+ __yend = 8 - __ystart;
+ }
+ } else { // INSIDE OF WINDOW
+ __ystart = 0;
+ __yend = 8;
+ }
+ if(((__ystart <= 0) && (__yend <= 0))) goto __noop;
+ for(int yy = (__ystart << 1); yy < (__yend << 1); yy += 2) {
__DECL_VECTORIZED_LOOP
for(int xx = 0; xx < 16; xx++) {
- lbuf[xx] = 0;
- mbuf[xx] = 0;
+ lbuf[xx].w = 0;
+ mbuf[xx].w = 0;
}
for(int yy2 = 0; yy2 < 2; yy2++) {
__DECL_VECTORIZED_LOOP
for(int xx = 0; xx < 16; xx++) {
- lbuf[xx] += (sbuf[yy + yy2][xx] & 0x7fff);
- mbuf[xx] |= (sbuf[yy + yy2][xx] & 0x8000);
+ lbuf[xx].w += (sbuf[yy + yy2][xx] & 0x7fff);
+ mbuf[xx].w |= (sbuf[yy + yy2][xx] & 0x8000);
}
}
__DECL_VECTORIZED_LOOP
for(int xx = 0; xx < 16; xx++) {
- lbuf[xx] = ((lbuf[xx] >> 1) & 0x7fff) | mbuf[xx];
+ lbuf[xx].w = ((lbuf[xx].w >> 1) & 0x7fff) | mbuf[xx].w;
+ }
+__DECL_VECTORIZED_LOOP
+ for(int xx = 0; xx < 16; xx++) {
+ mbuf[xx].w = (lbuf[xx].w == 0) ? 0xffff : 0x0000;
}
if(d_vram != NULL) {
- //d_vram->write_sprite_data(x, y + (yy >>1), xoffset, yoffset, lbuf, 16);
+ uint8_t* vp = d_vram->get_vram_address(vpaddr);
+ __DECL_ALIGNED(16) uint8_t source[32] = {0};
+ if(vp != NULL) {
+ for(int xx = (__xstart << 1); xx < (__xend << 1); xx++) {
+ source[xx] = vp[xx];
+ }
+__DECL_VECTORIZED_LOOP
+ for(int xx = 0; xx < 16; xx++) {
+ source[(xx << 1) + 0] &= mbuf[xx].b.l;
+ source[(xx << 1) + 1] &= mbuf[xx].b.h;
+ }
+__DECL_VECTORIZED_LOOP
+ for(int xx = 0; xx < 16; xx++) {
+ source[(xx << 1) + 0] |= lbuf[xx].b.l;
+ source[(xx << 1) + 1] |= lbuf[xx].b.h;
+ }
+ for(int xx = (__xstart << 1); xx < (__xend << 1); xx++) {
+ vp[xx] = source[xx];
+ }
+ }
}
+ vpaddr = (vpaddr + (256 << 1)) & 0x1ffff;
}
} else { //halfx &&halfy
- for(int yy = 0; yy < 16; yy += 2) {
+ int __xstart;
+ int __xend;
+ if(x < xoffset) {
+ __xstart = ((xoffset - x + 8) >= 0) ? (xoffset - x + 8) : 0;
+ __xend = ((xoffset - x + 8) >= 0) ? 8 : 0;
+ } else if(x > (xoffset + 256)) {
+ __xstart = x - (xoffset + 256);
+ if(__xstart >= 8) {
+ __xstart = 0;
+ __xend = 0;
+ } else {
+ __xstart = 0;
+ __xend = 8 - __xstart;
+ }
+ } else { // INSIDE OF WINDOW
+ __xstart = 0;
+ __xend = 8;
+ }
+ if(((__xstart <= 0) && (__xend <= 0))) goto __noop;
+ int __ystart;
+ int __yend;
+ if(y < yoffset) {
+ __ystart = ((yoffset - y + 8) >= 0) ? (yoffset - y + 8) : 0;
+ __yend = ((yoffset - y + 8) >= 0) ? 8 : 0;
+ } else if(y > (yoffset + 256)) {
+ __ystart = y - (yoffset + 256);
+ if(__ystart >= 8) {
+ __ystart = 0;
+ __yend = 0;
+ } else {
+ __ystart = 0;
+ __yend = 8 - __ystart;
+ }
+ } else { // INSIDE OF WINDOW
+ __ystart = 0;
+ __yend = 8;
+ }
+ if(((__ystart <= 0) && (__yend <= 0))) goto __noop;
+ for(int yy = (__ystart << 1); yy < (__yend << 1); yy += 2) {
__DECL_VECTORIZED_LOOP
for(int xx = 0; xx < 16; xx++) {
- lbuf[xx] = 0;
- mbuf[xx] = 0;
+ lbuf[xx].w = 0;
+ mbuf[xx].w = 0;
}
for(int yy2 = 0; yy2 < 2; yy2++) {
__DECL_VECTORIZED_LOOP
for(int xx = 0; xx < 16; xx += 2) {
- lbuf[xx >> 1] += (sbuf[yy + yy2][xx] & 0x7fff);
- mbuf[xx >> 1] |= (sbuf[yy + yy2][xx] & 0x8000);
+ lbuf[xx >> 1].w += (sbuf[yy + yy2][xx] & 0x7fff);
+ lbuf[xx >> 1].w += (sbuf[yy + yy2][xx + 1] & 0x7fff);
+ mbuf[xx >> 1].w |= (sbuf[yy + yy2][xx] & 0x8000);
+ mbuf[xx >> 1].w |= (sbuf[yy + yy2][xx + 1] & 0x8000);
}
+__DECL_VECTORIZED_LOOP
+ for(int xx = 0; xx < 8; xx++) {
+ lbuf[xx].w >>= 1;
+ }
+ }
+__DECL_VECTORIZED_LOOP
+ for(int xx = 0; xx < 8; xx++) {
+ lbuf[xx].w = ((lbuf[xx].w >> 2) & 0x7fff) | mbuf[xx].w;
}
__DECL_VECTORIZED_LOOP
for(int xx = 0; xx < 8; xx++) {
- lbuf[xx] = ((lbuf[xx] >> 2) & 0x7fff) | mbuf[xx];
+ mbuf[xx].w = (lbuf[xx].w == 0x0000) ? 0xffff : 0x0000;
}
if(d_vram != NULL) {
//d_vram->write_sprite_data(x, y + (yy >>1), xoffset, yoffset, lbuf, 8);
+ uint8_t* vp = d_vram->get_vram_address(vpaddr);
+ __DECL_ALIGNED(16) uint8_t source[16] = {0};
+ if(vp != NULL) {
+ for(int xx = (__xstart << 1); xx < (__xend << 1); xx++) {
+ source[xx] = vp[xx];
+ }
+__DECL_VECTORIZED_LOOP
+ for(int xx = 0; xx < 8; xx++) {
+ source[(xx << 1) + 0] &= mbuf[xx].b.l;
+ source[(xx << 1) + 1] &= mbuf[xx].b.h;
+ }
+__DECL_VECTORIZED_LOOP
+ for(int xx = 0; xx < 8; xx++) {
+ source[(xx << 1) + 0] |= lbuf[xx].b.l;
+ source[(xx << 1) + 1] |= lbuf[xx].b.h;
+ }
+ for(int xx = (__xstart << 1); xx < (__xend << 1); xx++) {
+ vp[xx] = source[xx];
+ }
+ }
}
+ vpaddr = (vpaddr + (256 << 1)) & 0x1ffff;
}
}
+__noop:
now_transferring = false;
}
OSDN Git Service
