mirror/RetroArch
1
0
Fork 0
mirror of https://github.com/libretro/RetroArch synced 2025-01-26 18:35:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

Optimize rgb1555 -> argb8888 conversion in SSE2.

Browse Source
This commit is contained in:
Themaister 2012-09-02 20:03:47 +02:00
parent 86b444cb97
commit b9630e2a2c
1 changed files with 63 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

63
gfx/scaler/pixconv.c
View File

@ -17,7 +17,69 @@
#include <stdint.h>
#include <stdlib.h>
#include <string.h>
#include <stdio.h>
#if defined(__SSE2__)
#include <emmintrin.h>
#endif
#if defined(__SSE2__)
void conv_0rgb1555_argb8888(void *output_, const void *input_,
int width, int height,
int out_stride, int in_stride)
{
const uint16_t *input = (const uint16_t*)input_;
uint32_t *output = (uint32_t*)output_;
__m128i pix_mask_r = _mm_set1_epi16(0x1f << 10);
__m128i pix_mask_gb = _mm_set1_epi16(0x1f << 5);
__m128i mul15_mid = _mm_set1_epi16(0x4200);
__m128i mul15_hi = _mm_set1_epi16(0x0210);
__m128i a = _mm_set1_epi16(0x00ff);
int max_width = width - 7;
for (int h = 0; h < height; h++, output += out_stride >> 2, input += in_stride >> 1)
{
int w;
for (w = 0; w < max_width; w += 8)
{
__m128i in = _mm_loadu_si128((const __m128i*)(input + w));
__m128i r = _mm_and_si128(in, pix_mask_r);
__m128i g = _mm_and_si128(in, pix_mask_gb);
__m128i b = _mm_and_si128(_mm_slli_epi16(in, 5), pix_mask_gb);
r = _mm_mulhi_epi16(r, mul15_hi);
g = _mm_mulhi_epi16(g, mul15_mid);
b = _mm_mulhi_epi16(b, mul15_mid);
__m128i res_lo_bg = _mm_unpacklo_epi8(b, g);
__m128i res_hi_bg = _mm_unpackhi_epi8(b, g);
__m128i res_lo_ra = _mm_unpacklo_epi8(r, a);
__m128i res_hi_ra = _mm_unpackhi_epi8(r, a);
__m128i res_lo = _mm_or_si128(res_lo_bg, _mm_slli_si128(res_lo_ra, 2));
__m128i res_hi = _mm_or_si128(res_hi_bg, _mm_slli_si128(res_hi_ra, 2));
_mm_storeu_si128((__m128i*)(output + w + 0), res_lo);
_mm_storeu_si128((__m128i*)(output + w + 4), res_hi);
}
for (; w < width; w++)
{
uint32_t col = input[w];
uint32_t r = (col >> 10) & 0x1f;
uint32_t g = (col >> 5) & 0x1f;
uint32_t b = (col >> 0) & 0x1f;
r = (r << 3) | (r >> 2);
g = (g << 3) | (g >> 2);
b = (b << 3) | (b >> 2);
output[w] = (0xff << 24) | (r << 16) | (g << 8) | (b << 0);
}
}
}
#else
void conv_0rgb1555_argb8888(void *output_, const void *input_,
int width, int height,
int out_stride, int in_stride)
@ -41,6 +103,7 @@ void conv_0rgb1555_argb8888(void *output_, const void *input_,
}
}
}
#endif
void conv_0rgb1555_bgr24(void *output_, const void *input_,
int width, int height,