From aabf2c5c4991f72b661681d6d19edb4e144836e6 Mon Sep 17 00:00:00 2001 From: Fabian Giesen Date: Thu, 18 Dec 2014 07:41:41 -0800 Subject: [PATCH] stb_image: SSE2 YCbCr->RGB kernel --- stb_image.h | 126 ++++++++++++++++++++++++++++++++++++++-------------- 1 file changed, 92 insertions(+), 34 deletions(-) diff --git a/stb_image.h b/stb_image.h index 8b9e982..fed480d 100644 --- a/stb_image.h +++ b/stb_image.h @@ -25,7 +25,7 @@ - decode from memory or through FILE (define STBI_NO_STDIO to remove code) - decode from arbitrary I/O callbacks - - overridable dequantizing-IDCT, YCbCr-to-RGB conversion (define STBI_SIMD) + - SIMD acceleration on x86/x64 Latest revisions: 1.48 (2014-12-14) fix incorrectly-named assert() @@ -326,20 +326,6 @@ STBIDEF char *stbi_zlib_decode_noheader_malloc(const char *buffer, int len, int STBIDEF int stbi_zlib_decode_noheader_buffer(char *obuffer, int olen, const char *ibuffer, int ilen); -// define faster low-level operations (typically SIMD support) -#ifdef STBI_SIMD -typedef void (*stbi_YCbCr_to_RGB_run)(stbi_uc *output, stbi_uc const *y, stbi_uc const *cb, stbi_uc const *cr, int count, int step); -// compute a conversion from YCbCr to RGB -// 'count' pixels -// write pixels to 'output'; each pixel is 'step' bytes (either 3 or 4; if 4, write '255' as 4th), order R,G,B -// y: Y input channel -// cb: Cb input channel; scale/biased to be 0..255 -// cr: Cr input channel; scale/biased to be 0..255 - -STBIDEF void stbi_install_YCbCr_to_RGB(stbi_YCbCr_to_RGB_run func); -#endif // STBI_SIMD - - #ifdef __cplusplus } #endif @@ -1074,6 +1060,7 @@ typedef struct // kernels void (*idct_block_kernel)(stbi_uc *out, int out_stride, short data[64]); + void (*YCbCr_to_RGB_kernel)(stbi_uc *out, const stbi_uc *y, const stbi_uc *pcb, const stbi_uc *pcr, int count, int step); } stbi__jpeg; static int stbi__build_huffman(stbi__huffman *h, int *count) @@ -1355,15 +1342,6 @@ stbi_inline static stbi_uc stbi__clamp(int x) t1 += p2+p4; \ t0 += p1+p3; -#ifdef STBI_SIMD -static unsigned short stbi__dq_ones[64] = { - 1,1,1,1, 1,1,1,1, 1,1,1,1, 1,1,1,1, - 1,1,1,1, 1,1,1,1, 1,1,1,1, 1,1,1,1, - 1,1,1,1, 1,1,1,1, 1,1,1,1, 1,1,1,1, - 1,1,1,1, 1,1,1,1, 1,1,1,1, 1,1,1,1, -}; -#endif - // .344 seconds on 3*anemones.jpg static void stbi__idct_block(stbi_uc *out, int out_stride, short data[64]) { @@ -2045,12 +2023,94 @@ static void stbi__YCbCr_to_RGB_row(stbi_uc *out, const stbi_uc *y, const stbi_uc } } -#ifdef STBI_SIMD -static stbi_YCbCr_to_RGB_run stbi__YCbCr_installed = stbi__YCbCr_to_RGB_row; - -STBIDEF void stbi_install_YCbCr_to_RGB(stbi_YCbCr_to_RGB_run func) +#ifdef STBI_SSE2 +static void stbi__YCbCr_to_RGB_sse2(stbi_uc *out, stbi_uc const *y, stbi_uc const *pcb, stbi_uc const *pcr, int count, int step) { - stbi__YCbCr_installed = func; + int i = 0; + + // step == 3 is pretty ugly on the final interleave, and i'm not convinced + // it's useful in practice (you wouldn't use it for textures, for example). + // so just accelerate step == 4 case. + // + // note: unlike the IDCT, this isn't bit-identical to the integer version. + if (step == 4) { + // this is a fairly straightforward implementation and not super-optimized. + __m128i signflip = _mm_set1_epi8(-0x80); + __m128i cr_const0 = _mm_set1_epi16((short) ( 1.40200f*4096.0f)); + __m128i cr_const1 = _mm_set1_epi16((short) (-0.71414f*4096.0f)); + __m128i cb_const0 = _mm_set1_epi16((short) (-0.34414f*4096.0f)); + __m128i cb_const1 = _mm_set1_epi16((short) ( 1.77200f*4096.0f)); + __m128i y_bias = _mm_set1_epi16(8); + __m128i xw = _mm_set1_epi16(255); + + for (; i+7 < count; i += 8) { + // load + __m128i zero = _mm_setzero_si128(); + __m128i y_bytes = _mm_loadl_epi64((__m128i *) (y+i)); + __m128i cr_bytes = _mm_loadl_epi64((__m128i *) (pcr+i)); + __m128i cb_bytes = _mm_loadl_epi64((__m128i *) (pcb+i)); + __m128i cr_bias = _mm_xor_si128(cr_bytes, signflip); // -128 + __m128i cb_bias = _mm_xor_si128(cb_bytes, signflip); // -128 + + // unpack to short (and left-shift cr, cb by 8) + __m128i yw = _mm_unpacklo_epi8(y_bytes, zero); + __m128i crw = _mm_unpacklo_epi8(_mm_setzero_si128(), cr_bias); + __m128i cbw = _mm_unpacklo_epi8(_mm_setzero_si128(), cb_bias); + + // color transform + __m128i yws = _mm_slli_epi16(yw, 4); + __m128i cr0 = _mm_mulhi_epi16(cr_const0, crw); + __m128i cb0 = _mm_mulhi_epi16(cb_const0, cbw); + __m128i ywb = _mm_add_epi16(yws, y_bias); + __m128i cb1 = _mm_mulhi_epi16(cbw, cb_const1); + __m128i cr1 = _mm_mulhi_epi16(crw, cr_const1); + __m128i rws = _mm_add_epi16(cr0, ywb); + __m128i gwt = _mm_add_epi16(cb0, ywb); + __m128i bws = _mm_add_epi16(ywb, cb1); + __m128i gws = _mm_add_epi16(gwt, cr1); + + // descale + __m128i rw = _mm_srai_epi16(rws, 4); + __m128i bw = _mm_srai_epi16(bws, 4); + __m128i gw = _mm_srai_epi16(gws, 4); + + // back to byte, set up for transpose + __m128i brb = _mm_packus_epi16(rw, bw); + __m128i gxb = _mm_packus_epi16(gw, xw); + + // transpose to interleave channels + __m128i t0 = _mm_unpacklo_epi8(brb, gxb); + __m128i t1 = _mm_unpackhi_epi8(brb, gxb); + __m128i o0 = _mm_unpacklo_epi16(t0, t1); + __m128i o1 = _mm_unpackhi_epi16(t0, t1); + + // store + _mm_storeu_si128((__m128i *) (out + 0), o0); + _mm_storeu_si128((__m128i *) (out + 16), o1); + out += 32; + } + } + + for (; i < count; ++i) { + int y_fixed = (y[i] << 16) + 32768; // rounding + int r,g,b; + int cr = pcr[i] - 128; + int cb = pcb[i] - 128; + r = y_fixed + cr*float2fixed(1.40200f); + g = y_fixed - cr*float2fixed(0.71414f) - cb*float2fixed(0.34414f); + b = y_fixed + cb*float2fixed(1.77200f); + r >>= 16; + g >>= 16; + b >>= 16; + if ((unsigned) r > 255) { if (r < 0) r = 0; else r = 255; } + if ((unsigned) g > 255) { if (g < 0) g = 0; else g = 255; } + if ((unsigned) b > 255) { if (b < 0) b = 0; else b = 255; } + out[0] = (stbi_uc)r; + out[1] = (stbi_uc)g; + out[2] = (stbi_uc)b; + out[3] = 255; + out += step; + } } #endif @@ -2058,10 +2118,12 @@ STBIDEF void stbi_install_YCbCr_to_RGB(stbi_YCbCr_to_RGB_run func) static void stbi__setup_jpeg(stbi__jpeg *j) { j->idct_block_kernel = stbi__idct_block; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_row; #ifdef STBI_SSE2 if (stbi__sse2_available()) { j->idct_block_kernel = stbi__idct_sse2; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_sse2; } #endif } @@ -2167,11 +2229,7 @@ static stbi_uc *load_jpeg_image(stbi__jpeg *z, int *out_x, int *out_y, int *comp if (n >= 3) { stbi_uc *y = coutput[0]; if (z->s->img_n == 3) { - #ifdef STBI_SIMD - stbi__YCbCr_installed(out, y, coutput[1], coutput[2], z->s->img_x, n); - #else - stbi__YCbCr_to_RGB_row(out, y, coutput[1], coutput[2], z->s->img_x, n); - #endif + z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); } else for (i=0; i < z->s->img_x; ++i) { out[0] = out[1] = out[2] = y[i];