diff --git a/Source/Core/Common/MathUtil.h b/Source/Core/Common/MathUtil.h index da12822..14b9309 100644 --- a/Source/Core/Common/MathUtil.h +++ b/Source/Core/Common/MathUtil.h @@ -150,7 +150,7 @@ float MathFloatVectorSum(const std::vector&); #define ROUND_DOWN(x, a) ((x) & ~((a) - 1)) // Rounds down. 0 -> undefined -inline u64 Log2(u64 val) +inline int Log2(u64 val) { #if defined(__GNUC__) return 63 - __builtin_clzll(val); @@ -161,7 +161,7 @@ inline u64 Log2(u64 val) return result; #else - u64 result = -1; + int result = -1; while (val != 0) { val >>= 1; diff --git a/Source/Core/VideoBackends/OGL/TextureConverter.cpp b/Source/Core/VideoBackends/OGL/TextureConverter.cpp index 753bd64..dbbb605 100644 --- a/Source/Core/VideoBackends/OGL/TextureConverter.cpp +++ b/Source/Core/VideoBackends/OGL/TextureConverter.cpp @@ -117,7 +117,7 @@ void CreatePrograms() " ivec2 uv = ivec2(gl_FragCoord.xy);\n" // We switch top/bottom here. TODO: move this to screen blit. " ivec2 ts = textureSize(samp9, 0);\n" - " vec4 c0 = texelFetch(samp9, ivec2(uv.x/2, ts.y-uv.y-1), 0);\n" + " vec4 c0 = texelFetch(samp9, ivec2(uv.x>>1, ts.y-uv.y-1), 0);\n" " float y = mix(c0.b, c0.r, (uv.x & 1) == 1);\n" " float yComp = 1.164 * (y - 0.0625);\n" " float uComp = c0.g - 0.5;\n" diff --git a/Source/Core/VideoCommon/TextureConversionShader.cpp b/Source/Core/VideoCommon/TextureConversionShader.cpp index a7db334..76f26c6 100644 --- a/Source/Core/VideoCommon/TextureConversionShader.cpp +++ b/Source/Core/VideoCommon/TextureConversionShader.cpp @@ -9,6 +9,8 @@ #include #endif +#include "Common/MathUtil.h" + #include "VideoCommon/BPMemory.h" #include "VideoCommon/RenderBase.h" #include "VideoCommon/TextureConversionShader.h" @@ -90,21 +92,17 @@ void WriteSwizzler(char*& p, u32 format, API_TYPE ApiType) " float2 uv0 = float2(0.0, 0.0);\n" ); - WRITE(p, " uv1.x = uv1.x * %d;\n", samples); + WRITE(p, " uv1.x = uv1.x << %d;\n", Log2(samples)); - WRITE(p, " int yl = uv1.y / %d;\n", blkH); - WRITE(p, " int yb = yl * %d;\n", blkH); - WRITE(p, " int yoff = uv1.y - yb;\n"); + WRITE(p, " int yb = uv1.y & ~(%d - 1);\n", blkH); + WRITE(p, " int yoff = uv1.y & (%d - 1);\n", blkH); WRITE(p, " int xp = uv1.x + yoff * position.z;\n"); - WRITE(p, " int xel = xp / %d;\n", samples == 1 ? factor : blkW); - WRITE(p, " int xb = xel / %d;\n", blkH); - WRITE(p, " int xoff = xel - xb * %d;\n", blkH); - WRITE(p, " int xl = uv1.x * %d / %d;\n", factor, blkW); - WRITE(p, " int xib = uv1.x * %d - xl * %d;\n", factor, blkW); - WRITE(p, " int halfxb = xb / %d;\n", factor); - - WRITE(p, " sampleUv.x = xib + halfxb * %d;\n", blkW); - WRITE(p, " sampleUv.y = yb + xoff;\n"); + WRITE(p, " int xel = xp >> %d;\n", Log2(samples == 1 ? factor : blkW)); + WRITE(p, " int xib = (uv1.x << %d) & (%d - 1);\n", factor-1, blkW); + + WRITE(p, " sampleUv.x = xib + (xel >> %d << %d);\n", Log2(blkH * factor), Log2(blkW)); + WRITE(p, " sampleUv.y = yb + (xel & (%d - 1));\n", blkH); + WRITE(p, " bool first = 0 == (uv1.x & 1);\n"); } void WriteSampleColor(char*& p, const char* colorComp, const char* dest, int xoffset, API_TYPE ApiType) @@ -373,8 +371,6 @@ void WriteRGBA8Encoder(char* p,API_TYPE ApiType) { WriteSwizzler(p, GX_TF_RGBA8, ApiType); - WRITE(p, " bool first = xb == (halfxb * 2);\n"); - WRITE(p, " float4 texSample;\n"); WRITE(p, " float4 color0;\n"); WRITE(p, " float4 color1;\n"); @@ -563,8 +559,6 @@ void WriteZ24Encoder(char* p, API_TYPE ApiType) { WriteSwizzler(p, GX_TF_Z24X8, ApiType); - WRITE(p, " bool first = xb == (halfxb * 2);\n"); - WRITE(p, " float depth0;\n"); WRITE(p, " float depth1;\n"); WRITE(p, " float3 expanded0;\n");