Update to XXH3 for CRC_OPT

projects/msvc/GLideN64.vcxproj

@@ -352,7 +352,6 @@ copy /Y "$(OutDir)$(TargetName).*" "$(Mupen64PluginsDir_x64)")</Command>
     <ClCompile Include="..\..\src\windows\ZilmarAPIImpl_windows.cpp">
       <ExcludedFromBuild Condition="'$(Configuration)'=='Debug_mupenplus' Or '$(Configuration)'=='Release_mupenplus'">true</ExcludedFromBuild>
     </ClCompile>
-    <ClCompile Include="..\..\src\xxHash\xxhash.c" />
     <ClCompile Include="..\..\src\ZilmarPluginAPI.cpp">
       <ExcludedFromBuild Condition="'$(Configuration)'=='Debug_mupenplus' Or '$(Configuration)'=='Release_mupenplus'">true</ExcludedFromBuild>
     </ClCompile>
@@ -496,4 +495,4 @@ copy /Y "$(OutDir)$(TargetName).*" "$(Mupen64PluginsDir_x64)")</Command>
   <Import Project="$(VCTargetsPath)\Microsoft.Cpp.targets" />
   <ImportGroup Label="ExtensionTargets">
   </ImportGroup>
-</Project>
+</Project>

projects/msvc/GLideN64.vcxproj.filters

@@ -305,9 +305,6 @@
     <ClCompile Include="..\..\src\CRC_OPT.cpp">
       <Filter>Source Files</Filter>
     </ClCompile>
-    <ClCompile Include="..\..\src\xxHash\xxhash.c">
-      <Filter>Source Files</Filter>
-    </ClCompile>
     <ClCompile Include="..\..\src\DebugDump.cpp">
       <Filter>Source Files</Filter>
     </ClCompile>
@@ -782,4 +779,4 @@
       <Filter>Header Files\Graphics\OpenGL\ThreadedOpenGL</Filter>
     </ClInclude>
   </ItemGroup>
-</Project>
+</Project>

src/CMakeLists.txt

@@ -346,7 +346,6 @@ elseif(CRC_OPT)
   )
   list(APPEND GLideN64_SOURCES
     CRC_OPT.cpp
-    xxHash/xxhash.c
   )
 elseif(CRC_NEON)
   list(REMOVE_ITEM GLideN64_SOURCES
@@ -354,7 +353,6 @@ elseif(CRC_NEON)
   )
   list(APPEND GLideN64_SOURCES
     Neon/CRC_OPT_NEON.cpp
-    xxHash/xxhash.c
   )
 endif(CRC_ARMV8)
 

src/CRC.h

@@ -4,5 +4,5 @@
 void CRC_Init();
 
 u32 CRC_Calculate_Strict( u32 crc, const void *buffer, u32 count );
-u32 CRC_Calculate( u32 crc, const void *buffer, u32 count );
-u32 CRC_CalculatePalette( u32 crc, const void *buffer, u32 count );
+u64 CRC_Calculate( u64 crc, const void *buffer, u32 count );
+u64 CRC_CalculatePalette( u64 crc, const void *buffer, u32 count );

src/CRC32.cpp

@@ -32,16 +32,17 @@ void CRC_Init()
 	}
 }
 
-u32 CRC_Calculate( u32 crc, const void * buffer, u32 count )
+u64 CRC_Calculate( u64 crc, const void * buffer, u32 count )
 {
 	u8 *p;
-	u32 orig = crc;
+	u32 crc32 = static_cast<u32>(crc);
+	u32 orig = crc32;
 
 	p = (u8*) buffer;
 	while (count--)
-		crc = (crc >> 8) ^ CRCTable[(crc & 0xFF) ^ *p++];
+		crc32 = (crc32 >> 8) ^ CRCTable[(crc32 & 0xFF) ^ *p++];
 
-	return crc ^ orig;
+	return crc32 ^ orig;
 }
 
 u32 CRC_Calculate_Strict( u32 crc, const void * buffer, u32 count )
@@ -49,18 +50,19 @@ u32 CRC_Calculate_Strict( u32 crc, const void * buffer, u32 count )
 	return CRC_Calculate(crc, buffer, count);
 }
 
-u32 CRC_CalculatePalette(u32 crc, const void * buffer, u32 count )
+u64 CRC_CalculatePalette(u64 crc, const void * buffer, u32 count )
 {
 	u8 *p;
-	u32 orig = crc;
+	u32 crc32 = static_cast<u32>(crc);
+	u32 orig = crc32;
 
 	p = (u8*) buffer;
 	while (count--) {
-		crc = (crc >> 8) ^ CRCTable[(crc & 0xFF) ^ *p++];
-		crc = (crc >> 8) ^ CRCTable[(crc & 0xFF) ^ *p++];
+		crc32 = (crc32 >> 8) ^ CRCTable[(crc32 & 0xFF) ^ *p++];
+		crc32 = (crc32 >> 8) ^ CRCTable[(crc32 & 0xFF) ^ *p++];
 
 		p += 6;
 	}
 
-	return crc ^ orig;
+	return crc32 ^ orig;
 }

src/CRC32_ARMV8.cpp

@@ -17,34 +17,35 @@ void CRC_Init()
 {
 }
 
-u32 CRC_Calculate( u32 crc, const void * buffer, u32 count )
+u64 CRC_Calculate( u64 crc, const void * buffer, u32 count )
 {
 	u8 *p;
-	u32 orig = crc;
+	u32 crc32 = static_cast<u32>(crc);
+	u32 orig = crc32;
 
 	p = (u8*) buffer;
 
 	// use eight byte crc intrinsic __crc32d if count is high enough.
 	// __crc32d, __crc32w, __crc32h and __crc32b use polynomial 0x04C11DB7
 	while (count >= 8) {
-		crc = __crc32d(crc, *((u64*)p));
+		crc32 = __crc32d(crc32, *((u64*)p));
 		p += 8;
 		count -= 8;
 	}
 	if (count >= 4) {
-		crc = __crc32w(crc, *((u32*)p));
+		crc32 = __crc32w(crc32, *((u32*)p));
 		p += 4;
 		count -= 4;
 	}
 	if (count >= 2) {
-		crc = __crc32h(crc, *((u16*)p));
+		crc32 = __crc32h(crc32, *((u16*)p));
 		p += 2;
 		count -= 2;
 	}
 	if (count == 1)
-		crc = __crc32b(crc, *p);
+		crc32 = __crc32b(crc32, *p);
 
-	return crc ^ orig;
+	return crc32 ^ orig;
 }
 
 u32 CRC_Calculate_Strict( u32 crc, const void * buffer, u32 count )
@@ -52,17 +53,18 @@ u32 CRC_Calculate_Strict( u32 crc, const void * buffer, u32 count )
 	return CRC_Calculate(crc, buffer, count);
 }
 
-u32 CRC_CalculatePalette(u32 crc, const void * buffer, u32 count )
+u64 CRC_CalculatePalette(u64 crc, const void * buffer, u32 count )
 {
 	u8 *p;
-	u32 orig = crc;
+	u32 crc32 = static_cast<u32>(crc);
+	u32 orig = crc32;
 
 	p = (u8*) buffer;
 	while (count--) {
 		// use two byte intrinsic __crc32h
-		crc = __crc32h(crc, *((u16*)p));
+		crc32 = __crc32h(crc32, *((u16*)p));
 		p += 8;
 	}
 
-	return crc ^ orig;
+	return crc32 ^ orig;
 }

src/CRC_OPT.cpp

@@ -1,5 +1,5 @@
 #include "CRC.h"
-#include "xxHash/xxhash.h"
+#include "xxHash/xxh3.h"
 
 #define CRC32_POLYNOMIAL     0x04C11DB7
 
@@ -45,16 +45,16 @@ u32 CRC_Calculate_Strict( u32 crc, const void * buffer, u32 count )
 	return crc ^ orig;
 }
 
-u32 CRC_Calculate( u32 crc, const void * buffer, u32 count )
+u64 CRC_Calculate( u64 crc, const void * buffer, u32 count )
 {
-	return XXH32(buffer, count, crc);
+	return XXH3_64bits_withSeed(buffer, count, crc);
 }
 
-u32 CRC_CalculatePalette(u32 crc, const void * buffer, u32 count )
+u64 CRC_CalculatePalette( u64 crc, const void * buffer, u32 count )
 {
 	u8 *p = (u8*) buffer;
 	while (count--) {
-		crc = XXH32(p, 2, crc);
+		crc = XXH3_64bits_withSeed(p, 2, crc);
 		p += 8;
 	}
 	return crc;

src/Graphics/OpenGLContext/opengl_BufferedDrawer.cpp

@@ -109,7 +109,7 @@ void BufferedDrawer::_updateRectBuffer(const graphics::Context::DrawRectParamete
 		return;
 	}
 
-	const u32 crc = CRC_Calculate(0xFFFFFFFF, _params.vertices, dataSize);
+	const u64 crc = CRC_Calculate(UINT64_MAX, _params.vertices, dataSize);
 	auto iter = m_rectBufferOffsets.find(crc);
 	if (iter != m_rectBufferOffsets.end()) {
 		buffer.pos = iter->second;

src/Graphics/OpenGLContext/opengl_BufferedDrawer.h

@@ -74,7 +74,7 @@ namespace opengl {
 
 		std::vector<Vertex> m_vertices;
 
-		typedef std::unordered_map<u32, u32> BufferOffsets;
+		typedef std::unordered_map<u64, u32> BufferOffsets;
 		BufferOffsets m_rectBufferOffsets;
 
 		static const u32 m_bufMaxSize;

src/Neon/CRC_OPT_NEON.cpp

@@ -119,11 +119,11 @@ u32 ReliableHash32NEON(const void *input, size_t len, u32 seed) {
 	return h32;
 }
 
-u32 CRC_Calculate(u32 crc, const void *buffer, u32 count) {
+u64 CRC_Calculate(u64 crc, const void *buffer, u32 count) {
 	return ReliableHash32NEON(buffer, count, crc);
 }
 
-u32 CRC_CalculatePalette(u32 crc, const void *buffer, u32 count) {
+u64 CRC_CalculatePalette(u64 crc, const void *buffer, u32 count) {
 	u8 *p = (u8 *) buffer;
 	while (count--) {
 		crc = ReliableHash32NEON(p, 2, crc);

src/PaletteTexture.h

@@ -15,7 +15,7 @@ public:
 private:
 	CachedTexture * m_pTexture;
 	u8* m_pbuf;
-	u32 m_paletteCRC256;
+	u64 m_paletteCRC256;
 };
 
 extern PaletteTexture g_paletteTexture;

src/Textures.cpp

@@ -525,15 +525,15 @@ void TextureCache::_checkCacheSize()
 	}
 }
 
-CachedTexture * TextureCache::_addTexture(u32 _crc32)
+CachedTexture * TextureCache::_addTexture(u64 _crc64)
 {
 	if (m_curUnpackAlignment == 0)
 		m_curUnpackAlignment = gfxContext.getTextureUnpackAlignment();
 	_checkCacheSize();
 	m_textures.emplace_front(gfxContext.createTexture(textureTarget::TEXTURE_2D));
 	Textures::iterator new_iter = m_textures.begin();
-	new_iter->crc = _crc32;
-	m_lruTextureLocations.insert(std::pair<u32, Textures::iterator>(_crc32, new_iter));
+	new_iter->crc = _crc64;
+	m_lruTextureLocations.insert(std::pair<u64, Textures::iterator>(_crc64, new_iter));
 	return &(*new_iter);
 }
 
@@ -1241,7 +1241,7 @@ struct TextureParams
 };
 
 static
-u32 _calculateCRC(u32 _t, const TextureParams & _params, u32 _bytes)
+u64 _calculateCRC(u32 _t, const TextureParams & _params, u32 _bytes)
 {
 	const bool rgba32 = gSP.textureTile[_t]->size == G_IM_SIZ_32b;
 	if (_bytes == 0) {
@@ -1252,7 +1252,7 @@ u32 _calculateCRC(u32 _t, const TextureParams & _params, u32 _bytes)
 		_bytes >>= 1;
 	const u32 tMemMask = (gDP.otherMode.textureLUT == G_TT_NONE && !rgba32) ? 0x1FF : 0xFF;
 	const u64 *src = (u64*)&TMEM[gSP.textureTile[_t]->tmem & tMemMask];
-	u32 crc = 0xFFFFFFFF;
+	u64 crc = UINT64_MAX;
 	crc = CRC_Calculate(crc, src, _bytes);
 
 	if (rgba32) {
@@ -1350,9 +1350,9 @@ void TextureCache::activateMSDummy(u32 _t)
 void TextureCache::_updateBackground()
 {
 	u32 numBytes = gSP.bgImage.width * gSP.bgImage.height << gSP.bgImage.size >> 1;
-	u32 crc;
+	u64 crc;
 
-	crc = CRC_Calculate( 0xFFFFFFFF, &RDRAM[gSP.bgImage.address], numBytes );
+	crc = CRC_Calculate( UINT64_MAX, &RDRAM[gSP.bgImage.address], numBytes );
 
 	if (gDP.otherMode.textureLUT != G_TT_NONE || gSP.bgImage.format == G_IM_FMT_CI) {
 		if (gSP.bgImage.size == G_IM_SIZ_4b)
@@ -1502,7 +1502,7 @@ void TextureCache::update(u32 _t)
 	params.width = sizes.width;
 	params.height = sizes.height;
 
-	const u32 crc = _calculateCRC(_t, params, sizes.bytes);
+	const u64 crc = _calculateCRC(_t, params, sizes.bytes);
 
 	if (current[_t] != nullptr && current[_t]->crc == crc) {
 		activateTexture(_t, current[_t]);

src/Textures.h

@@ -18,7 +18,7 @@ struct CachedTexture
 	CachedTexture(graphics::ObjectHandle _name) : name(_name), max_level(0), frameBufferTexture(fbNone), bHDTexture(false) {}
 
 	graphics::ObjectHandle name;
-	u32		crc = 0;
+	u64		crc = 0;
 //	float	fulS, fulT;
 //	WORD	ulS, ulT, lrS, lrT;
 	float	offsetS, offsetT;
@@ -78,7 +78,7 @@ private:
 	TextureCache(const TextureCache &) = delete;
 
 	void _checkCacheSize();
-	CachedTexture * _addTexture(u32 _crc32);
+	CachedTexture * _addTexture(u64 _crc64);
 	void _load(u32 _tile, CachedTexture *_pTexture);
 	bool _loadHiresTexture(u32 _tile, CachedTexture *_pTexture, u64 & _ricecrc);
 	void _loadBackground(CachedTexture *pTexture);
@@ -90,7 +90,7 @@ private:
 	void _getTextureDestData(CachedTexture& tmptex, u32* pDest, graphics::Parameter glInternalFormat, GetTexelFunc GetTexel, u16* pLine);
 
 	typedef std::list<CachedTexture> Textures;
-	typedef std::unordered_map<u32, Textures::iterator> Texture_Locations;
+	typedef std::unordered_map<u64, Textures::iterator> Texture_Locations;
 	typedef std::unordered_map<u32, CachedTexture> FBTextures;
 	Textures m_textures;
 	Texture_Locations m_lruTextureLocations;

src/gDP.cpp

@@ -708,11 +708,11 @@ void gDPLoadTLUT( u32 tile, u32 uls, u32 ult, u32 lrs, u32 lrt )
 			destIdx += 4;
 		}
 
-		gDP.paletteCRC16[pal] = CRC_CalculatePalette(0xFFFFFFFF, &TMEM[256 + (pal << 4)], 16);
+		gDP.paletteCRC16[pal] = CRC_CalculatePalette(UINT64_MAX, &TMEM[256 + (pal << 4)], 16);
 		pal = (pal + 1) & 0x0F;
 	}
 
-	gDP.paletteCRC256 = CRC_Calculate(0xFFFFFFFF, gDP.paletteCRC16, 64);
+	gDP.paletteCRC256 = CRC_Calculate(UINT64_MAX, gDP.paletteCRC16, 64);
 
 	if (TFH.isInited()) {
 		const u16 start = gDP.tiles[tile].tmem - 256; // starting location in the palettes

src/gDP.h

@@ -251,8 +251,8 @@ struct gDPInfo
 	u32 changed;
 
 	u16 TexFilterPalette[512];
-	u32 paletteCRC16[16];
-	u32 paletteCRC256;
+	u64 paletteCRC16[16];
+	u64 paletteCRC256;
 	u32 half_1, half_2;
 
 	 gDPLoadTileInfo loadInfo[512];