(view as text)
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<float>&);
#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/DolphinWX/GameListCtrl.cpp b/Source/Core/DolphinWX/GameListCtrl.cpp
index bd420f8..6029c74 100644
--- a/Source/Core/DolphinWX/GameListCtrl.cpp
+++ b/Source/Core/DolphinWX/GameListCtrl.cpp
@@ -413,12 +413,12 @@ wxString NiceSizeFormat(u64 _size)
{
const char* const unit_symbols[] = {"B", "KiB", "MiB", "GiB", "TiB", "PiB", "EiB", "ZiB", "YiB"};
- auto const unit = Log2(std::max<u64>(_size, 1)) / 10;
- auto const unit_size = (1 << (unit * 10));
+ const u64 unit = Log2(std::max<u64>(_size, 1)) / 10;
+ const u64 unit_size = (1 << (unit * 10));
// ugly rounding integer math
- auto const value = (_size + unit_size / 2) / unit_size;
- auto const frac = (_size % unit_size * 10 + unit_size / 2) / unit_size % 10;
+ const u64 value = (_size + unit_size / 2) / unit_size;
+ const u64 frac = (_size % unit_size * 10 + unit_size / 2) / unit_size % 10;
return StrToWxStr(StringFromFormat("%" PRIu64 ".%" PRIu64 " %s", value, frac, unit_symbols[unit]));
}
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/TextureCacheBase.cpp b/Source/Core/VideoCommon/TextureCacheBase.cpp
index 5abce6f..d48c03c 100644
--- a/Source/Core/VideoCommon/TextureCacheBase.cpp
+++ b/Source/Core/VideoCommon/TextureCacheBase.cpp
@@ -886,4 +886,7 @@ void TextureCache::CopyRenderTargetToTexture(u32 dstAddr, unsigned int dstFormat
entry->frameCount = frameCount;
entry->FromRenderTarget(dstAddr, dstFormat, srcFormat, srcRect, isIntensity, scaleByHalf, cbufid, colmat);
+
+ delete entry;
+ textures.erase(dstAddr);
}
diff --git a/Source/Core/VideoCommon/TextureConversionShader.cpp b/Source/Core/VideoCommon/TextureConversionShader.cpp
index a7db334..9054eac 100644
--- a/Source/Core/VideoCommon/TextureConversionShader.cpp
+++ b/Source/Core/VideoCommon/TextureConversionShader.cpp
@@ -14,6 +14,7 @@
#include "VideoCommon/TextureConversionShader.h"
#include "VideoCommon/TextureDecoder.h"
#include "VideoCommon/VideoConfig.h"
+#include "Common/MathUtil.h"
#define WRITE p+=sprintf
@@ -65,8 +66,7 @@ void WriteSwizzler(char*& p, u32 format, API_TYPE ApiType)
int blkW = TexDecoder_GetBlockWidthInTexels(format);
int blkH = TexDecoder_GetBlockHeightInTexels(format);
int samples = GetEncodedSampleCount(format);
- // 32 bit textures (RGBA8 and Z24) are store in 2 cache line increments
- int factor = samples == 1 ? 2 : 1;
+
if (ApiType == API_OPENGL)
{
WRITE(p, "#define samp0 samp9\n");
@@ -90,21 +90,21 @@ 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, " 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 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 y_block_position = uv1.y & ~(%d - 1);\n", blkH);
+ WRITE(p, " int y_offset_in_block = uv1.y & (%d - 1);\n", blkH);
+ WRITE(p, " int x_virtual_position = (uv1.x << %d) + y_offset_in_block * position.z;\n", Log2(samples));
+ WRITE(p, " int x_block_position = (x_virtual_position >> %d) & ~(%d - 1);\n", Log2(blkH), blkW);
+ if (samples == 1)
+ {
+ // 32 bit textures (RGBA8 and Z24) are store in 2 cache line increments
+ WRITE(p, " x_virtual_position = x_virtual_position << 1;\n");
+ WRITE(p, " bool first = 0 == (uv1.x & 8);\n"); // first cache line, used in the encoders
+ }
+ WRITE(p, " int x_offset_in_block = x_virtual_position & (%d - 1);\n", blkW);
+ WRITE(p, " int y_offset = (x_virtual_position >> %d) & (%d - 1);\n", Log2(blkW), blkH);
+
+ WRITE(p, " sampleUv.x = x_offset_in_block + x_block_position;\n");
+ WRITE(p, " sampleUv.y = y_block_position + y_offset;\n");
}
void WriteSampleColor(char*& p, const char* colorComp, const char* dest, int xoffset, API_TYPE ApiType)
@@ -373,8 +373,6 @@ 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 +561,6 @@ 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");