Sprinter/mame
2
0
Fork 0
mirror of https://github.com/holub/mame synced 2025-04-24 01:11:11 +03:00
Code Issues Actions Packages Projects Releases Wiki Activity

- Added SIMD code for the rest of the RSP vector load instructions [MooglyGuy]

Browse Source
This commit is contained in:
Ryan Holtz 2013-06-09 08:41:21 +00:00
parent 3a2aabb05d
commit fa4cca37ba
1 changed files with 78 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

78
src/emu/cpu/rsp/rspdrc.c
View File

@ -973,11 +973,28 @@ static void cfunc_rsp_lrv(void *param)
end = 16;
ea &= ~0xf;
#if USE_SIMD
INT16 mask[8] = { 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000 };
INT16 val[8] = { 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000 };
#endif
for (i=index; i < end; i++)
{
#if USE_SIMD
mask[i >> 1] |= 0x00ff << ((i & 1) * 8);
val[i >> 1] |= READ8(rsp, ea) << ((i & 1) * 8);
#endif
VREG_B(dest, i) = READ8(rsp, ea);
ea++;
}
#if USE_SIMD
__m128i neg1 = _mm_set_epi16(0xffff, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff);
__m128i keep_mask = _mm_set_epi16(mask[0], mask[1], mask[2], mask[3], mask[4], mask[5], mask[6], mask[7]);
__m128i load_val = _mm_set_epi16(val[0], val[1], val[2], val[3], val[4], val[5], val[6], val[7]);
keep_mask = _mm_xor_si128(keep_mask, neg1);
rsp->xv[dest] = _mm_and_si128(rsp->xv[dest], keep_mask);
rsp->xv[dest] = _mm_or_si128(rsp->xv[dest], load_val);
#endif
}
static void cfunc_rsp_lpv(void *param)
@ -1003,10 +1020,20 @@ static void cfunc_rsp_lpv(void *param)
ea = (base) ? rsp->r[base] + (offset * 8) : (offset * 8);
#if USE_SIMD
INT16 val[8] = { 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000 };
#endif
for (i=0; i < 8; i++)
{
#if USE_SIMD
val[i] = READ8(rsp, ea + (((16-index) + i) & 0xf)) << 8;
#endif
W_VREG_S(dest, i) = READ8(rsp, ea + (((16-index) + i) & 0xf)) << 8;
}
#if USE_SIMD
rsp->xv[dest] = _mm_set_epi16(val[0], val[1], val[2], val[3], val[4], val[5], val[6], val[7]);
#endif
}
static void cfunc_rsp_luv(void *param)
@ -1032,10 +1059,20 @@ static void cfunc_rsp_luv(void *param)
ea = (base) ? rsp->r[base] + (offset * 8) : (offset * 8);
#if USE_SIMD
INT16 val[8] = { 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000 };
#endif
for (i=0; i < 8; i++)
{
#if USE_SIMD
val[i] = READ8(rsp, ea + (((16-index) + i) & 0xf)) << 7;
#endif
W_VREG_S(dest, i) = READ8(rsp, ea + (((16-index) + i) & 0xf)) << 7;
}
#if USE_SIMD
rsp->xv[dest] = _mm_set_epi16(val[0], val[1], val[2], val[3], val[4], val[5], val[6], val[7]);
#endif
}
static void cfunc_rsp_lhv(void *param)
@ -1061,10 +1098,20 @@ static void cfunc_rsp_lhv(void *param)
ea = (base) ? rsp->r[base] + (offset * 16) : (offset * 16);
#if USE_SIMD
INT16 val[8] = { 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000 };
#endif
for (i=0; i < 8; i++)
{
#if USE_SIMD
val[i] = READ8(rsp, ea + (((16-index) + (i<<1)) & 0xf)) << 7;
#endif
W_VREG_S(dest, i) = READ8(rsp, ea + (((16-index) + (i<<1)) & 0xf)) << 7;
}
#if USE_SIMD
rsp->xv[dest] = _mm_set_epi16(val[0], val[1], val[2], val[3], val[4], val[5], val[6], val[7]);
#endif
}
static void cfunc_rsp_lfv(void *param)
@ -1096,11 +1143,28 @@ static void cfunc_rsp_lfv(void *param)
end = (index >> 1) + 4;
#if USE_SIMD
INT16 mask[8] = { 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000 };
INT16 val[8] = { 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000 };
#endif
for (i=index >> 1; i < end; i++)
{
#if USE_SIMD
mask[i] = 0xffff;
val[i] = READ8(rsp, ea) << 7;
#endif
W_VREG_S(dest, i) = READ8(rsp, ea) << 7;
ea += 4;
}
#if USE_SIMD
__m128i neg1 = _mm_set_epi16(0xffff, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff);
__m128i keep_mask = _mm_set_epi16(mask[0], mask[1], mask[2], mask[3], mask[4], mask[5], mask[6], mask[7]);
__m128i load_val = _mm_set_epi16(val[0], val[1], val[2], val[3], val[4], val[5], val[6], val[7]);
keep_mask = _mm_xor_si128(keep_mask, neg1);
rsp->xv[dest] = _mm_and_si128(rsp->xv[dest], keep_mask);
rsp->xv[dest] = _mm_or_si128(rsp->xv[dest], load_val);
#endif
}
static void cfunc_rsp_lwv(void *param)
@ -1130,11 +1194,21 @@ static void cfunc_rsp_lwv(void *param)
end = (16 - index) + 16;
#if USE_SIMD
INT16 val[8] = { 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000 };
#endif
for (i=(16 - index); i < end; i++)
{
#if USE_SIMD
val[i >> 1] |= READ8(rsp, ea) << ((i & 1) * 8);
#endif
VREG_B(dest, i & 0xf) = READ8(rsp, ea);
ea += 4;
}
#if USE_SIMD
rsp->xv[dest] = _mm_set_epi16(val[0], val[1], val[2], val[3], val[4], val[5], val[6], val[7]);
#endif
}
static void cfunc_rsp_ltv(void *param)
@ -1173,6 +1247,10 @@ static void cfunc_rsp_ltv(void *param)
for (i = vs; i < ve; i++)
{
element = ((8 - (index >> 1) + (i - vs)) << 1);
#if USE_SIMD
UINT16 value = (READ8(rsp, ea + 1) << 8) | READ8(rsp, ea);
_mm_insert_epi16 (rsp->xv[i], value, element);
#endif
VREG_B(i, (element & 0xf)) = READ8(rsp, ea);
VREG_B(i, ((element + 1) & 0xf)) = READ8(rsp, ea + 1);