diff --git a/CMakeLists.txt b/CMakeLists.txt
index 4182f0342d..2225d7fc34 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -7,6 +7,19 @@ cmake_minimum_required(VERSION 2.6)
 # Variable to check that people use the good file
 set(TOP_CMAKE_WAS_SOURCED TRUE)
 
+# Print a clear message that 64bits is not supported
+# It would avoid compilation failure later.
+if(CMAKE_SIZEOF_VOID_P MATCHES "8")
+    message(FATAL_ERROR "
+    PCSX2 does not support 64bits environment. Please install a 32bits chroot or a 32bits OS.
+    PCSX2 have neither no plan to support the 64bits architecture in the future.
+    It will need a complete rewrite of the core emulator and a lots of time
+    
+    However when linux distribution will support properly multi-arch package, it will 
+    be at least possible to easily compile and install PCSX2 witout too much hassle (the chroot environment)")
+endif(CMAKE_SIZEOF_VOID_P MATCHES "8")
+
+
 # set module path
 set(CMAKE_MODULE_PATH ${PROJECT_SOURCE_DIR}/cmake)
 
diff --git a/common/src/Utilities/x86/MemcpyVibes.cpp b/common/src/Utilities/x86/MemcpyVibes.cpp
index 32d13895a1..b154cd2847 100644
--- a/common/src/Utilities/x86/MemcpyVibes.cpp
+++ b/common/src/Utilities/x86/MemcpyVibes.cpp
@@ -98,8 +98,8 @@ void gen_memcpy_vibes() {
 			xADD(ecx, 256);
 		}
 		const xRegisterSSE xmm_t(x);
-		xMOVAPS(xmm_t, ptr32[edx+off]);
-		xMOVAPS(ptr32[ecx+off], xmm_t);
+		xMOVAPS (xmm_t, ptr32[edx+off]);
+		xMOVNTPS(ptr32[ecx+off], xmm_t);
 	}
 
 	_memcpy_vibes[0] = (_memCpyCall)xGetPtr();
diff --git a/pcsx2/IPU/yuv2rgb.cpp b/pcsx2/IPU/yuv2rgb.cpp
index 423cacc71c..5fcda2ef18 100644
--- a/pcsx2/IPU/yuv2rgb.cpp
+++ b/pcsx2/IPU/yuv2rgb.cpp
@@ -63,11 +63,11 @@ struct SSE2_Tables
 	u16 Y_mask[8];			// offset -32
 	u16 round_1bit[8];		// offset -16
 
-	u16 Y_coefficients[8];	// offset 0
-	u16 GCr_coefficients[8];// offset 16
-	u16 GCb_coefficients[8];// offset 32
-	u16 RCr_coefficients[8];// offset 48
-	u16 BCb_coefficients[8];// offset 64
+	s16 Y_coefficients[8];	// offset 0
+	s16 GCr_coefficients[8];// offset 16
+	s16 GCb_coefficients[8];// offset 32
+	s16 RCr_coefficients[8];// offset 48
+	s16 BCb_coefficients[8];// offset 64
 };
 
 enum
diff --git a/pcsx2/x86/newVif_UnpackSSE.cpp b/pcsx2/x86/newVif_UnpackSSE.cpp
index 7e089b9e7b..32cc3c1f0a 100644
--- a/pcsx2/x86/newVif_UnpackSSE.cpp
+++ b/pcsx2/x86/newVif_UnpackSSE.cpp
@@ -98,27 +98,31 @@ void VifUnpackSSE_Base::xUPK_S_32() const {
 }
 
 void VifUnpackSSE_Base::xUPK_S_16() const {
-if (x86caps.hasStreamingSIMD4Extensions) {
-	xPMOVXX16  (workReg);
-}
-else {
-	xMOV16     (workReg, ptr32[srcIndirect]);
-	xPUNPCK.LWD(workReg, workReg);
-	xShiftR    (workReg, 16);
-}
+	if (x86caps.hasStreamingSIMD4Extensions) 
+	{
+		xPMOVXX16  (workReg);
+	}
+	else 
+	{
+		xMOV16     (workReg, ptr32[srcIndirect]);
+		xPUNPCK.LWD(workReg, workReg);
+		xShiftR    (workReg, 16);
+	}
 	xPSHUF.D   (destReg, workReg, _v0);
 }
 
 void VifUnpackSSE_Base::xUPK_S_8() const {
-if (x86caps.hasStreamingSIMD4Extensions) {
-	xPMOVXX8   (workReg);
-}
-else {
-	xMOV8      (workReg, ptr32[srcIndirect]);
-	xPUNPCK.LBW(workReg, workReg);
-	xPUNPCK.LWD(workReg, workReg);
-	xShiftR    (workReg, 24);
-}
+	if (x86caps.hasStreamingSIMD4Extensions) 
+	{
+		xPMOVXX8   (workReg);
+	}
+	else 
+	{
+		xMOV8      (workReg, ptr32[srcIndirect]);
+		xPUNPCK.LBW(workReg, workReg);
+		xPUNPCK.LWD(workReg, workReg);
+		xShiftR    (workReg, 24);
+	}
 	xPSHUF.D   (destReg, workReg, _v0);
 }
 
@@ -133,58 +137,63 @@ void VifUnpackSSE_Base::xUPK_V2_32() const {
 }
 
 void VifUnpackSSE_Base::xUPK_V2_16() const {
-if (x86caps.hasStreamingSIMD4Extensions) {
-	xPMOVXX16  (destReg);
-}
-else {
-	xMOV32     (destReg, ptr32[srcIndirect]);
-	xPUNPCK.LWD(destReg, destReg);
-	xShiftR    (destReg, 16);
-}
+	if (x86caps.hasStreamingSIMD4Extensions) 
+	{
+		xPMOVXX16  (destReg);
+	}
+	else 
+	{
+		xMOV32     (destReg, ptr32[srcIndirect]);
+		xPUNPCK.LWD(destReg, destReg);
+		xShiftR    (destReg, 16);
+	}
 	xPSHUF.D   (destReg, destReg, 0x44); //v1v0v1v0
 }
 
 void VifUnpackSSE_Base::xUPK_V2_8() const {
-if (x86caps.hasStreamingSIMD4Extensions) {
-	xPMOVXX8   (destReg);
-}
-else {
-	xMOV16     (destReg, ptr32[srcIndirect]);
-	xPUNPCK.LBW(destReg, destReg);
-	xPUNPCK.LWD(destReg, destReg);
-	xShiftR    (destReg, 24);
-}
+	if (x86caps.hasStreamingSIMD4Extensions) 
+	{
+		xPMOVXX8   (destReg);
+	}
+	else 
+	{
+		xMOV16     (destReg, ptr32[srcIndirect]);
+		xPUNPCK.LBW(destReg, destReg);
+		xPUNPCK.LWD(destReg, destReg);
+		xShiftR    (destReg, 24);
+	}
 	xPSHUF.D   (destReg, destReg, 0x44); //v1v0v1v0
 }
 
 void VifUnpackSSE_Base::xUPK_V3_32() const {
 	xMOV128    (destReg, ptr32[srcIndirect]);
-	xPSHUF.D   (destReg, destReg, 0xA4); //v2v2v1v0
 }
 
 void VifUnpackSSE_Base::xUPK_V3_16() const {
-if (x86caps.hasStreamingSIMD4Extensions) {
-	xPMOVXX16  (destReg);
-}
-else {
-	xMOV64     (destReg, ptr32[srcIndirect]);
-	xPUNPCK.LWD(destReg, destReg);
-	xShiftR    (destReg, 16);
-}
-	xPSHUF.D   (destReg, destReg, 0xA4); //v2v2v1v0
+	if (x86caps.hasStreamingSIMD4Extensions) 
+	{
+		xPMOVXX16  (destReg);
+	}
+	else 
+	{
+		xMOV64     (destReg, ptr32[srcIndirect]);
+		xPUNPCK.LWD(destReg, destReg);
+		xShiftR    (destReg, 16);
+	}
 }
 
 void VifUnpackSSE_Base::xUPK_V3_8() const {
-if (x86caps.hasStreamingSIMD4Extensions) {
-	xPMOVXX8   (destReg);
-}
-else {
-	xMOV32     (destReg, ptr32[srcIndirect]);
-	xPUNPCK.LBW(destReg, destReg);
-	xPUNPCK.LWD(destReg, destReg);
-	xShiftR    (destReg, 24);
-}
-	xPSHUF.D   (destReg, destReg, 0xA4); //v2v2v1v0
+	if (x86caps.hasStreamingSIMD4Extensions) 
+	{
+		xPMOVXX8   (destReg);
+	}
+	else 
+	{
+		xMOV32     (destReg, ptr32[srcIndirect]);
+		xPUNPCK.LBW(destReg, destReg);
+		xPUNPCK.LWD(destReg, destReg);
+		xShiftR    (destReg, 24);
+	}
 }
 
 void VifUnpackSSE_Base::xUPK_V4_32() const {
@@ -192,26 +201,30 @@ void VifUnpackSSE_Base::xUPK_V4_32() const {
 }
 
 void VifUnpackSSE_Base::xUPK_V4_16() const {
-if (x86caps.hasStreamingSIMD4Extensions) {
-	xPMOVXX16  (destReg);
-}
-else {
-	xMOV64     (destReg, ptr32[srcIndirect]);
-	xPUNPCK.LWD(destReg, destReg);
-	xShiftR    (destReg, 16);
-}
+	if (x86caps.hasStreamingSIMD4Extensions) 
+	{
+		xPMOVXX16  (destReg);
+	}
+	else 
+	{
+		xMOV64     (destReg, ptr32[srcIndirect]);
+		xPUNPCK.LWD(destReg, destReg);
+		xShiftR    (destReg, 16);
+	}
 }
 
 void VifUnpackSSE_Base::xUPK_V4_8() const {
-if (x86caps.hasStreamingSIMD4Extensions) {
-	xPMOVXX8   (destReg);
-}
-else {
-	xMOV32     (destReg, ptr32[srcIndirect]);
-	xPUNPCK.LBW(destReg, destReg);
-	xPUNPCK.LWD(destReg, destReg);
-	xShiftR    (destReg, 24);
-}
+	if (x86caps.hasStreamingSIMD4Extensions) 
+	{
+		xPMOVXX8   (destReg);
+	}
+	else 
+	{
+		xMOV32     (destReg, ptr32[srcIndirect]);
+		xPUNPCK.LBW(destReg, destReg);
+		xPUNPCK.LWD(destReg, destReg);
+		xShiftR    (destReg, 24);
+	}
 }
 
 void VifUnpackSSE_Base::xUPK_V4_5() const {
diff --git a/plugins/GSdx/GSCapture.cpp b/plugins/GSdx/GSCapture.cpp
index b2207e09fb..de8ee02dea 100644
--- a/plugins/GSdx/GSCapture.cpp
+++ b/plugins/GSdx/GSCapture.cpp
@@ -245,53 +245,31 @@ public:
 		{
 			int dstpitch = ((VIDEOINFOHEADER*)mt.Format())->bmiHeader.biWidth * 2;
 
-			const GSVector4 ys(0.098f, 0.504f, 0.257f, 0.0f);
-			const GSVector4 us(0.439f / 2, -0.291f / 2, -0.148f / 2, 0.0f);
-			const GSVector4 vs(-0.071f / 2, -0.368f / 2, 0.439f / 2, 0.0f);
+			GSVector4 ys(0.257f, 0.504f, 0.098f, 0.0f);
+			GSVector4 us(-0.148f / 2, -0.291f / 2, 0.439f / 2, 0.0f);
+			GSVector4 vs(0.439f / 2, -0.368f / 2, -0.071f / 2, 0.0f);
 			const GSVector4 offset(16, 128, 16, 128);
 
-			if(rgba)
+			if (!rgba)
+				ys = ys.zyxw(), us = us.zyxw(), vs = vs.zyxw();
+
+			for(int j = 0; j < h; j++, dst += dstpitch, src += srcpitch)
 			{
-				for(int j = 0; j < h; j++, dst += dstpitch, src += srcpitch)
+				uint32* s = (uint32*)src;
+				uint16* d = (uint16*)dst;
+
+				for(int i = 0; i < w; i += 2)
 				{
-					uint32* s = (uint32*)src;
-					uint16* d = (uint16*)dst;
+					GSVector4 c0 = GSVector4(s[i + 0]);
+					GSVector4 c1 = GSVector4(s[i + 1]);
+					GSVector4 c2 = c0 + c1;
 
-					for(int i = 0; i < w; i += 2)
-					{
-						GSVector4 c0 = GSVector4(s[i + 0]);
-						GSVector4 c1 = GSVector4(s[i + 1]);
-						GSVector4 c2 = c0 + c1;
+					GSVector4 lo = (c0 * ys).hadd(c2 * us);
+					GSVector4 hi = (c1 * ys).hadd(c2 * vs);
 
-						GSVector4 lo = (c0 * ys).hadd(c2 * vs);
-						GSVector4 hi = (c1 * ys).hadd(c2 * us);
+					GSVector4 c = lo.hadd(hi) + offset;
 
-						GSVector4 c = lo.hadd(hi) + offset;
-
-						*((uint32*)&d[i]) = GSVector4i(c).rgba32();
-					}
-				}
-			}
-			else
-			{
-				for(int j = 0; j < h; j++, dst += dstpitch, src += srcpitch)
-				{
-					uint32* s = (uint32*)src;
-					uint16* d = (uint16*)dst;
-
-					for(int i = 0; i < w; i += 2)
-					{
-						GSVector4 c0 = GSVector4(s[i + 0]).zyxw();
-						GSVector4 c1 = GSVector4(s[i + 1]).zyxw();
-						GSVector4 c2 = c0 + c1;
-
-						GSVector4 lo = (c0 * ys).hadd(c2 * vs);
-						GSVector4 hi = (c1 * ys).hadd(c2 * us);
-
-						GSVector4 c = lo.hadd(hi) + offset;
-
-						*((uint32*)&d[i]) = GSVector4i(c).rgba32();
-					}
+					*((uint32*)&d[i]) = GSVector4i(c).rgba32();
 				}
 			}
 		}
diff --git a/plugins/zzogl-pg/opengl/GSmain.cpp b/plugins/zzogl-pg/opengl/GSmain.cpp
index 9e15dc7d85..dd46a08ee3 100644
--- a/plugins/zzogl-pg/opengl/GSmain.cpp
+++ b/plugins/zzogl-pg/opengl/GSmain.cpp
@@ -126,6 +126,53 @@ void CALLBACK GSsetLogDir(const char* dir)
 	ZZLog::SetDir(dir);
 }
 
+void ReportHacks(gameHacks hacks)
+{
+	if (hacks.texture_targs) ZZLog::WriteLn("'Texture targs' hack enabled.");
+	if (hacks.auto_reset) ZZLog::WriteLn("'Auto reset' hack enabled.");
+	if (hacks.interlace_2x) ZZLog::WriteLn("'Interlace 2x' hack enabled.");
+	if (hacks.texa) ZZLog::WriteLn("'Texa' hack enabled.");
+	if (hacks.no_target_resolve) ZZLog::WriteLn("'No target resolve' hack enabled.");
+	if (hacks.exact_color) ZZLog::WriteLn("Exact color hack enabled.");
+	if (hacks.no_color_clamp) ZZLog::WriteLn("'No color clamp' hack enabled.");
+	if (hacks.no_alpha_fail) ZZLog::WriteLn("'No alpha fail' hack enabled.");
+	if (hacks.no_depth_update) ZZLog::WriteLn("'No depth update' hack enabled.");
+	if (hacks.quick_resolve_1) ZZLog::WriteLn("'Quick resolve 1' enabled.");
+	if (hacks.no_quick_resolve) ZZLog::WriteLn("'No Quick resolve' hack enabled.");
+	if (hacks.no_target_clut) ZZLog::WriteLn("'No target clut' hack enabled.");
+	if (hacks.vss_hack_off) ZZLog::WriteLn("VSS hack enabled.");
+	if (hacks.no_depth_resolve) ZZLog::WriteLn("'No depth resolve' hack enabled.");
+	if (hacks.full_16_bit_res) ZZLog::WriteLn("'Full 16 bit resolution' hack enabled.");
+	if (hacks.resolve_promoted) ZZLog::WriteLn("'Resolve promoted' hack enabled.");
+	if (hacks.fast_update) ZZLog::WriteLn("'Fast update' hack enabled.");
+	if (hacks.no_alpha_test) ZZLog::WriteLn("'No alpha test' hack enabled.");
+	if (hacks.disable_mrt_depth) ZZLog::WriteLn("'Disable mrt depth' hack enabled.");
+	if (hacks.args_32_bit) ZZLog::WriteLn("'Args 32 bit' hack enabled.");
+	if (hacks.path3) ZZLog::WriteLn("'Path3' hack enabled.");
+	if (hacks.parallel_context) ZZLog::WriteLn("'Parallel context' hack enabled.");
+	if (hacks.xenosaga_spec) ZZLog::WriteLn("'Xenosaga spec' hack enabled.");
+	if (hacks.partial_pointers) ZZLog::WriteLn("'Partial pointers' hack enabled.");
+	if (hacks.partial_depth) ZZLog::WriteLn("'Partial depth' hack enabled.");
+	if (hacks.reget) ZZLog::WriteLn("Reget hack enabled.");
+	if (hacks.gust) ZZLog::WriteLn("Gust hack enabled.");
+	if (hacks.no_logz) ZZLog::WriteLn("'No logz' hack enabled.");
+}
+
+void ListHacks()
+{
+	if (conf.def_hacks._u32 != 0)
+	{
+		ZZLog::WriteLn("AutoEnabling these hacks:");
+		ReportHacks(conf.def_hacks);
+	}
+	
+	if (conf.hacks._u32 != 0)
+	{
+		ZZLog::WriteLn("You've manually enabled these hacks:");
+		ReportHacks(conf.hacks);
+	}
+}
+
 void CALLBACK GSsetGameCRC(int crc, int options)
 {
 	// TEXDESTROY_THRESH starts out at 16.
@@ -133,15 +180,15 @@ void CALLBACK GSsetGameCRC(int crc, int options)
 	conf.mrtdepth = (conf.settings().disable_mrt_depth != 0);
 
 	if (!conf.mrtdepth)
-		ZZLog::Error_Log("Disabling MRT depth writing.");
+		ZZLog::WriteLn("Disabling MRT depth writing.");
 	else
-		ZZLog::Error_Log("Enabling MRT depth writing.");
+		ZZLog::WriteLn("Enabling MRT depth writing.");
 
 	bool CRCValueChanged = (g_LastCRC != crc);
 
 	g_LastCRC = crc;
 
-	ZZLog::Error_Log("CRC = %x", crc);
+	if (crc != 0) ZZLog::WriteLn("Current game CRC is %x.", crc);
 
 	if (CRCValueChanged && (crc != 0))
 	{
@@ -149,17 +196,27 @@ void CALLBACK GSsetGameCRC(int crc, int options)
 		{
 			if (crc_game_list[i].crc == crc)
 			{
-				if (crc_game_list[i].v_thresh > 0) VALIDATE_THRESH = crc_game_list[i].v_thresh;
-				if (crc_game_list[i].t_thresh > 0) TEXDESTROY_THRESH = crc_game_list[i].t_thresh;
+				ZZLog::WriteLn("Found CRC[%x] in crc game list.", crc);
+				
+				if (crc_game_list[i].v_thresh > 0) 
+				{
+					VALIDATE_THRESH = crc_game_list[i].v_thresh;
+					ZZLog::WriteLn("Setting VALIDATE_THRESH to %d", VALIDATE_THRESH);
+				}
+				
+				if (crc_game_list[i].t_thresh > 0) 
+				{
+					TEXDESTROY_THRESH = crc_game_list[i].t_thresh;
+					ZZLog::WriteLn("Setting TEXDESTROY_THRESH to %d", TEXDESTROY_THRESH);
+				}
 
 				conf.def_hacks._u32 |= crc_game_list[i].flags;
-
-				ZZLog::Error_Log("Found CRC[%x] in crc game list.", crc);
-
+				ListHacks();
 				return;
 			}
 		}
 	}
+	ListHacks();
 }
 
 void CALLBACK GSsetFrameSkip(int frameskip)
diff --git a/plugins/zzogl-pg/opengl/HostMemory.cpp b/plugins/zzogl-pg/opengl/HostMemory.cpp
index d8074035c3..2f45eab73d 100644
--- a/plugins/zzogl-pg/opengl/HostMemory.cpp
+++ b/plugins/zzogl-pg/opengl/HostMemory.cpp
@@ -27,7 +27,6 @@
 #include "zerogs.h"
 #include "targets.h"
 
-
  namespace ZeroGS
  {
 	extern CRangeManager s_RangeMngr; // manages overwritten memory
@@ -249,8 +248,10 @@
 	}
 
 	template <class T>
-	void TransferLocalHost(void* pbyMem, u32 nQWordSize, int& x, int& y, u8 *pstart, _readPixel_0 rp)
+	void TransferLocalHost(void* pbyMem, u32 nQWordSize, int& x, int& y, u8 *pstart)
 	{
+		_readPixel_0 rp = readPixelFun_0[gs.srcbuf.psm];
+		
 		int i = x, j = y;
 		T* pbuf = (T*)pbyMem;
 		u32 nSize = nQWordSize * 16 / sizeof(T);
@@ -275,8 +276,10 @@
 		}
 	}
 
-	void TransferLocalHost_24(void* pbyMem, u32 nQWordSize, int& x, int& y, u8 *pstart, _readPixel_0 rp)
+	void TransferLocalHost_24(void* pbyMem, u32 nQWordSize, int& x, int& y, u8 *pstart)
 	{
+		_readPixel_0 rp = readPixelFun_0[gs.srcbuf.psm];
+		
 		int i = x, j = y;
 		u8* pbuf = (u8*)pbyMem;
 		u32 nSize = nQWordSize * 16 / 3;
@@ -312,65 +315,113 @@
 		assert(gs.imageTransfer == 1);
 
 		u8* pstart = g_pbyGSMemory + 256 * gs.srcbuf.bp;
-		int i = gs.imageY, j = gs.imageX;
-
-		switch (gs.srcbuf.psm)
+		
+		switch(PSMT_BITMODE(gs.srcbuf.psm))
 		{
-
-			case PSMCT32:
-				TransferLocalHost<u32>(pbyMem, nQWordSize, i, j, pstart, readPixel32_0);
-				break;
-
-			case PSMCT24:
-				TransferLocalHost_24(pbyMem, nQWordSize, i, j, pstart, readPixel24_0);
-				break;
-
-			case PSMCT16:
-				TransferLocalHost<u16>(pbyMem, nQWordSize, i, j, pstart, readPixel16_0);
-				break;
-
-			case PSMCT16S:
-				TransferLocalHost<u16>(pbyMem, nQWordSize, i, j, pstart, readPixel16S_0);
-				break;
-
-			case PSMT8:
-				TransferLocalHost<u8>(pbyMem, nQWordSize, i, j, pstart, readPixel8_0);
-				break;
-
-			case PSMT8H:
-				TransferLocalHost<u8>(pbyMem, nQWordSize, i, j, pstart, readPixel8H_0);
-				break;
-
-			case PSMT32Z:
-				TransferLocalHost<u32>(pbyMem, nQWordSize, i, j, pstart, readPixel32Z_0);
-				break;
-
-			case PSMT24Z:
-				TransferLocalHost_24(pbyMem, nQWordSize, i, j, pstart, readPixel24Z_0);
-				break;
-
-			case PSMT16Z:
-				TransferLocalHost<u16>(pbyMem, nQWordSize, i, j, pstart, readPixel16Z_0);
-				break;
-
-			case PSMT16SZ:
-				TransferLocalHost<u16>(pbyMem, nQWordSize, i, j, pstart, readPixel16SZ_0);
-				break;
-
-			default:
-				assert(0);
+			case 0: TransferLocalHost<u32>(pbyMem, nQWordSize, gs.imageY, gs.imageX, pstart); break;
+			case 1: TransferLocalHost_24(pbyMem, nQWordSize, gs.imageY, gs.imageX, pstart); break;
+			case 2: TransferLocalHost<u16>(pbyMem, nQWordSize, gs.imageY, gs.imageX, pstart); break;
+			case 3: TransferLocalHost<u8>(pbyMem, nQWordSize, gs.imageY, gs.imageX, pstart); break;
+			default: assert(0); break;
 		}
 
-		gs.imageY = i;
-		gs.imageX = j;
-
 		if (gs.imageY >= gs.imageEndY)
 		{
+			ZZLog::Error_Log("gs.imageY >= gs.imageEndY!");
 			assert(gs.imageY == gs.imageEndY);
 			gs.imageTransfer = -1;
 		}
 	}
+	
+__forceinline void _TransferLocalLocal()
+{
+	//ZZLog::Error_Log("TransferLocalLocal(0x%x, 0x%x)", gs.srcbuf.psm, gs.dstbuf.psm);
+	_writePixel_0 wp = writePixelFun_0[gs.srcbuf.psm];
+	_readPixel_0 rp = readPixelFun_0[gs.dstbuf.psm];
+	u8* pSrcBuf = g_pbyGSMemory + gs.srcbuf.bp * 256;
+	u8* pDstBuf = g_pbyGSMemory + gs.dstbuf.bp * 256;
+	u32 widthlimit = 4;
+	u32 maxX = gs.trxpos.sx + gs.imageWnew;
+	u32 maxY = gs.trxpos.sy + gs.imageHnew;
+	
+	if (PSMT_BITMODE(gs.srcbuf.psm) == 0) widthlimit = 2;
+	if ((gs.imageWnew & widthlimit) != 0) return;
+	
+	for(int i = gs.trxpos.sy, i2 = gs.trxpos.dy; i < maxY; i++, i2++) 
+	{
+		for(int j = gs.trxpos.sx, j2 = gs.trxpos.dx; j < maxX; j += widthlimit, j2 += widthlimit)
+		{
+			wp(pDstBuf, j2%2048, i2%2048,
+				rp(pSrcBuf, j%2048, i%2048, gs.srcbuf.bw), gs.dstbuf.bw);
+			
+			wp(pDstBuf, (j2+1)%2048, i2%2048,
+				rp(pSrcBuf, (j+1)%2048, i%2048, gs.srcbuf.bw), gs.dstbuf.bw);
+				
+			if (widthlimit > 2) 
+			{
+				// Then widthlimit == 4.
+				wp(pDstBuf, (j2+2)%2048, i2%2048,
+					rp(pSrcBuf, (j+2)%2048, i%2048, gs.srcbuf.bw), gs.dstbuf.bw);
+						
+				wp(pDstBuf, (j2+3)%2048, i2%2048,
+					rp(pSrcBuf, (j+3)%2048, i%2048, gs.srcbuf.bw), gs.dstbuf.bw);
+			}
+		}
+	}
+}
 
+__forceinline void _TransferLocalLocal_4()
+{
+	//ZZLog::Error_Log("TransferLocalLocal_4(0x%x, 0x%x)", gs.srcbuf.psm, gs.dstbuf.psm);
+	_getPixelAddress_0 gsp = getPixelFun_0[gs.srcbuf.psm];
+	_getPixelAddress_0 gdp = getPixelFun_0[gs.dstbuf.psm];
+	u8* pSrcBuf = g_pbyGSMemory + gs.srcbuf.bp * 256;
+	u8* pDstBuf = g_pbyGSMemory + gs.dstbuf.bp * 256;
+	u32 maxX = gs.trxpos.sx + gs.imageWnew;
+	u32 maxY = gs.trxpos.sy + gs.imageHnew;
+
+	assert((gs.imageWnew % 8) == 0);
+		
+	for(int i = gs.trxpos.sy, i2 = gs.trxpos.dy; i < maxY; ++i, ++i2) 
+	{
+		for(int j = gs.trxpos.sx, j2 = gs.trxpos.dx; j < maxX; j += 8, j2 += 8) 
+		{
+			/* NOTE: the 2 conseq 4bit values are in NOT in the same byte */
+			u32 read = gsp(j%2048, i%2048, gs.srcbuf.bw);
+			u32 write = gdp(j2%2048, i2%2048, gs.dstbuf.bw);
+			pDstBuf[write] = (pDstBuf[write]&0xf0)|(pSrcBuf[read]&0x0f);
+	
+			read = gsp((j+1)%2048, i%2048, gs.srcbuf.bw);
+			write = gdp((j2+1)%2048, i2%2048, gs.dstbuf.bw);
+			pDstBuf[write] = (pDstBuf[write]&0x0f)|(pSrcBuf[read]&0xf0);
+	
+			read = gsp((j+2)%2048, i%2048, gs.srcbuf.bw);
+			write = gdp((j2+2)%2048, i2%2048, gs.dstbuf.bw);
+			pDstBuf[write] = (pDstBuf[write]&0xf0)|(pSrcBuf[read]&0x0f);
+	
+			read = gsp((j+3)%2048, i%2048, gs.srcbuf.bw);
+			write = gdp((j2+3)%2048, i2%2048, gs.dstbuf.bw);
+			pDstBuf[write] = (pDstBuf[write]&0x0f)|(pSrcBuf[read]&0xf0);
+	
+			read = gsp((j+2)%2048, i%2048, gs.srcbuf.bw);
+			write = gdp((j2+2)%2048, i2%2048, gs.dstbuf.bw);
+			pDstBuf[write] = (pDstBuf[write]&0xf0)|(pSrcBuf[read]&0x0f);
+	
+			read = gsp((j+3)%2048, i%2048, gs.srcbuf.bw);
+			write = gdp((j2+3)%2048, i2%2048, gs.dstbuf.bw);
+			pDstBuf[write] = (pDstBuf[write]&0x0f)|(pSrcBuf[read]&0xf0);
+	
+			read = gsp((j+2)%2048, i%2048, gs.srcbuf.bw);
+			write = gdp((j2+2)%2048, i2%2048, gs.dstbuf.bw);
+			pDstBuf[write] = (pDstBuf[write]&0xf0)|(pSrcBuf[read]&0x0f);
+	
+			read = gsp((j+3)%2048, i%2048, gs.srcbuf.bw);
+			write = gdp((j2+3)%2048, i2%2048, gs.dstbuf.bw);
+			pDstBuf[write] = (pDstBuf[write]&0x0f)|(pSrcBuf[read]&0xf0);
+		}
+	}
+}
+	
 	// dir depends on trxpos.dirx & trxpos.diry
 	void TransferLocalLocal()
 	{
@@ -408,285 +459,19 @@
 				//(*it)->status |= CRenderTarget::TS_NeedUpdate;
 			}
 		}
-
-		u8* pSrcBuf = g_pbyGSMemory + gs.srcbuf.bp * 256;
-		u8* pDstBuf = g_pbyGSMemory + gs.dstbuf.bp * 256;
-
-	#define TRANSFERLOCALLOCAL(srcpsm, dstpsm, widthlimit) { \
-		if( (gs.imageWnew&widthlimit)!=0 ) break; \
-		assert( (gs.imageWnew&widthlimit)==0 && widthlimit <= 4); \
-		for(int i = gs.trxpos.sy, i2 = gs.trxpos.dy; i < gs.trxpos.sy+gs.imageHnew; i++, i2++) { \
-			for(int j = gs.trxpos.sx, j2 = gs.trxpos.dx; j < gs.trxpos.sx+gs.imageWnew; j+=widthlimit, j2+=widthlimit) { \
-				\
-				writePixel##dstpsm##_0(pDstBuf, j2%2048, i2%2048, \
-					readPixel##srcpsm##_0(pSrcBuf, j%2048, i%2048, gs.srcbuf.bw), gs.dstbuf.bw); \
-				\
-				if( widthlimit > 1 ) { \
-					writePixel##dstpsm##_0(pDstBuf, (j2+1)%2048, i2%2048, \
-						readPixel##srcpsm##_0(pSrcBuf, (j+1)%2048, i%2048, gs.srcbuf.bw), gs.dstbuf.bw); \
-					\
-					if( widthlimit > 2 ) { \
-						writePixel##dstpsm##_0(pDstBuf, (j2+2)%2048, i2%2048, \
-							readPixel##srcpsm##_0(pSrcBuf, (j+2)%2048, i%2048, gs.srcbuf.bw), gs.dstbuf.bw); \
-						\
-						if( widthlimit > 3 ) { \
-							writePixel##dstpsm##_0(pDstBuf, (j2+3)%2048, i2%2048, \
-								readPixel##srcpsm##_0(pSrcBuf, (j+3)%2048, i%2048, gs.srcbuf.bw), gs.dstbuf.bw); \
-						} \
-					} \
-				} \
-			} \
-		} \
-	} \
 	 
-	#define TRANSFERLOCALLOCAL_4(srcpsm, dstpsm) { \
-		assert( (gs.imageWnew%8) == 0 ); \
-		for(int i = gs.trxpos.sy, i2 = gs.trxpos.dy; i < gs.trxpos.sy+gs.imageHnew; ++i, ++i2) { \
-			for(int j = gs.trxpos.sx, j2 = gs.trxpos.dx; j < gs.trxpos.sx+gs.imageWnew; j+=8, j2+=8) { \
-				/* NOTE: the 2 conseq 4bit values are in NOT in the same byte */ \
-				u32 read = getPixelAddress##srcpsm##_0(j%2048, i%2048, gs.srcbuf.bw); \
-				u32 write = getPixelAddress##dstpsm##_0(j2%2048, i2%2048, gs.dstbuf.bw); \
-				pDstBuf[write] = (pDstBuf[write]&0xf0)|(pSrcBuf[read]&0x0f); \
-	 \
-				read = getPixelAddress##srcpsm##_0((j+1)%2048, i%2048, gs.srcbuf.bw); \
-				write = getPixelAddress##dstpsm##_0((j2+1)%2048, i2%2048, gs.dstbuf.bw); \
-				pDstBuf[write] = (pDstBuf[write]&0x0f)|(pSrcBuf[read]&0xf0); \
-	 \
-				read = getPixelAddress##srcpsm##_0((j+2)%2048, i%2048, gs.srcbuf.bw); \
-				write = getPixelAddress##dstpsm##_0((j2+2)%2048, i2%2048, gs.dstbuf.bw); \
-				pDstBuf[write] = (pDstBuf[write]&0xf0)|(pSrcBuf[read]&0x0f); \
-	 \
-				read = getPixelAddress##srcpsm##_0((j+3)%2048, i%2048, gs.srcbuf.bw); \
-				write = getPixelAddress##dstpsm##_0((j2+3)%2048, i2%2048, gs.dstbuf.bw); \
-				pDstBuf[write] = (pDstBuf[write]&0x0f)|(pSrcBuf[read]&0xf0); \
-	 \
-				read = getPixelAddress##srcpsm##_0((j+2)%2048, i%2048, gs.srcbuf.bw); \
-				write = getPixelAddress##dstpsm##_0((j2+2)%2048, i2%2048, gs.dstbuf.bw); \
-				pDstBuf[write] = (pDstBuf[write]&0xf0)|(pSrcBuf[read]&0x0f); \
-	 \
-				read = getPixelAddress##srcpsm##_0((j+3)%2048, i%2048, gs.srcbuf.bw); \
-				write = getPixelAddress##dstpsm##_0((j2+3)%2048, i2%2048, gs.dstbuf.bw); \
-				pDstBuf[write] = (pDstBuf[write]&0x0f)|(pSrcBuf[read]&0xf0); \
-	 \
-				read = getPixelAddress##srcpsm##_0((j+2)%2048, i%2048, gs.srcbuf.bw); \
-				write = getPixelAddress##dstpsm##_0((j2+2)%2048, i2%2048, gs.dstbuf.bw); \
-				pDstBuf[write] = (pDstBuf[write]&0xf0)|(pSrcBuf[read]&0x0f); \
-	 \
-				read = getPixelAddress##srcpsm##_0((j+3)%2048, i%2048, gs.srcbuf.bw); \
-				write = getPixelAddress##dstpsm##_0((j2+3)%2048, i2%2048, gs.dstbuf.bw); \
-				pDstBuf[write] = (pDstBuf[write]&0x0f)|(pSrcBuf[read]&0xf0); \
-			} \
-		} \
-	} \
-	 
-		switch (gs.srcbuf.psm) 
+		if (PSMT_BITMODE(gs.srcbuf.psm) != 4)
 		{
-			case PSMCT32:
-				if (gs.dstbuf.psm == PSMCT32)
-				{
-					TRANSFERLOCALLOCAL(32, 32, 2);
-				}
-				else
-				{
-					TRANSFERLOCALLOCAL(32, 32Z, 2);
-				}
-				break;
-
-			case PSMCT24:
-				if (gs.dstbuf.psm == PSMCT24)
-				{
-					TRANSFERLOCALLOCAL(24, 24, 4);
-				}
-				else
-				{
-					TRANSFERLOCALLOCAL(24, 24Z, 4);
-				}
-				break;
-
-			case PSMCT16:
-				switch (gs.dstbuf.psm)
-				{
-					case PSMCT16:
-						TRANSFERLOCALLOCAL(16, 16, 4);
-						break;
-
-					case PSMCT16S:
-						TRANSFERLOCALLOCAL(16, 16S, 4);
-						break;
-
-					case PSMT16Z:
-						TRANSFERLOCALLOCAL(16, 16Z, 4);
-						break;
-
-					case PSMT16SZ:
-						TRANSFERLOCALLOCAL(16, 16SZ, 4);
-						break;
-				}
-				break;
-
-			case PSMCT16S:
-				switch (gs.dstbuf.psm)
-				{
-					case PSMCT16:
-						TRANSFERLOCALLOCAL(16S, 16, 4);
-						break;
-
-					case PSMCT16S:
-						TRANSFERLOCALLOCAL(16S, 16S, 4);
-						break;
-
-					case PSMT16Z:
-						TRANSFERLOCALLOCAL(16S, 16Z, 4);
-						break;
-
-					case PSMT16SZ:
-						TRANSFERLOCALLOCAL(16S, 16SZ, 4);
-						break;
-				}
-				break;
-
-			case PSMT8:
-				if (gs.dstbuf.psm == PSMT8)
-				{
-					TRANSFERLOCALLOCAL(8, 8, 4);
-				}
-				else
-				{
-					TRANSFERLOCALLOCAL(8, 8H, 4);
-				}
-				break;
-
-			case PSMT4:
-				switch (gs.dstbuf.psm)
-				{
-
-					case PSMT4:
-						TRANSFERLOCALLOCAL_4(4, 4);
-						break;
-
-					case PSMT4HL:
-						TRANSFERLOCALLOCAL_4(4, 4HL);
-						break;
-
-					case PSMT4HH:
-						TRANSFERLOCALLOCAL_4(4, 4HH);
-						break;
-				}
-				break;
-
-			case PSMT8H:
-				if (gs.dstbuf.psm == PSMT8)
-				{
-					TRANSFERLOCALLOCAL(8H, 8, 4);
-				}
-				else
-				{
-					TRANSFERLOCALLOCAL(8H, 8H, 4);
-				}
-				break;
-
-			case PSMT4HL:
-				switch (gs.dstbuf.psm)
-				{
-					case PSMT4:
-						TRANSFERLOCALLOCAL_4(4HL, 4);
-						break;
-
-					case PSMT4HL:
-						TRANSFERLOCALLOCAL_4(4HL, 4HL);
-						break;
-
-					case PSMT4HH:
-						TRANSFERLOCALLOCAL_4(4HL, 4HH);
-						break;
-				}
-				break;
-
-			case PSMT4HH:
-				switch (gs.dstbuf.psm)
-				{
-					case PSMT4:
-						TRANSFERLOCALLOCAL_4(4HH, 4);
-						break;
-
-					case PSMT4HL:
-						TRANSFERLOCALLOCAL_4(4HH, 4HL);
-						break;
-
-					case PSMT4HH:
-						TRANSFERLOCALLOCAL_4(4HH, 4HH);
-						break;
-				}
-				break;
-
-			case PSMT32Z:
-				if (gs.dstbuf.psm == PSMCT32)
-				{
-					TRANSFERLOCALLOCAL(32Z, 32, 2);
-				}
-				else
-				{
-					TRANSFERLOCALLOCAL(32Z, 32Z, 2);
-				}
-				break;
-
-			case PSMT24Z:
-				if (gs.dstbuf.psm == PSMCT24)
-				{
-					TRANSFERLOCALLOCAL(24Z, 24, 4);
-				}
-				else
-				{
-					TRANSFERLOCALLOCAL(24Z, 24Z, 4);
-				}
-				break;
-
-			case PSMT16Z:
-				switch (gs.dstbuf.psm)
-				{
-					case PSMCT16:
-						TRANSFERLOCALLOCAL(16Z, 16, 4);
-						break;
-
-					case PSMCT16S:
-						TRANSFERLOCALLOCAL(16Z, 16S, 4);
-						break;
-
-					case PSMT16Z:
-						TRANSFERLOCALLOCAL(16Z, 16Z, 4);
-						break;
-
-					case PSMT16SZ:
-						TRANSFERLOCALLOCAL(16Z, 16SZ, 4);
-						break;
-				}
-				break;
-
-			case PSMT16SZ:
-				switch (gs.dstbuf.psm)
-				{
-					case PSMCT16:
-						TRANSFERLOCALLOCAL(16SZ, 16, 4);
-						break;
-
-					case PSMCT16S:
-						TRANSFERLOCALLOCAL(16SZ, 16S, 4);
-						break;
-
-					case PSMT16Z:
-						TRANSFERLOCALLOCAL(16SZ, 16Z, 4);
-						break;
-
-					case PSMT16SZ:
-						TRANSFERLOCALLOCAL(16SZ, 16SZ, 4);
-						break;
-				}
-				break;
+			_TransferLocalLocal();
+		}
+		else
+		{
+			_TransferLocalLocal_4();
 		}
 
 		g_MemTargs.ClearRange(dststart, dstend);
 
-	#ifdef DEVBUILD
+	#ifdef ZEROGS_DEVBUILD
 
 		if (g_bSaveTrans)
 		{
diff --git a/plugins/zzogl-pg/opengl/Mem.h b/plugins/zzogl-pg/opengl/Mem.h
index 2cc7627c58..7b347078fc 100644
--- a/plugins/zzogl-pg/opengl/Mem.h
+++ b/plugins/zzogl-pg/opengl/Mem.h
@@ -44,6 +44,10 @@ typedef int (*_TransferHostLocal)(const void* pbyMem, u32 nQWordSize);
 typedef void (*_TransferLocalHost)(void* pbyMem, u32 nQWordSize);
 typedef void (*_SwizzleBlock)(u8 *dst, u8 *src, int pitch, u32 WriteMask);
 
+extern _getPixelAddress_0 getPixelFun_0[64];
+extern _writePixel_0 writePixelFun_0[64];
+extern _readPixel_0 readPixelFun_0[64];
+
 enum Psm_Size
 {
 	PSM_ = 0,
@@ -269,8 +273,6 @@ static __forceinline u32 getPixelAddress16SZ_0(int x, int y, u32 bw)
 #define getPixelAddress_0(psm,x,y,bw) getPixelAddress##psm##_0(x,y,bw)
 #define getPixelAddress(psm,x,y,bp,bw) getPixelAddress##psm##(x,y,bp,bw)
 
-
-
 static __forceinline void writePixel32(void* pmem, int x, int y, u32 pixel, u32 bp, u32 bw)
 {
 	((u32*)pmem)[getPixelAddress32(x, y, bp, bw)] = pixel;
@@ -511,7 +513,6 @@ static __forceinline void writePixel16SZ_0(void* pmem, int x, int y, u32 pixel,
 	((u16*)pmem)[getPixelAddress16SZ_0(x, y, bw)] = pixel;
 }
 
-
 ///////////////
 
 static __forceinline u32 readPixel32_0(const void* pmem, int x, int y, u32 bw)
diff --git a/plugins/zzogl-pg/opengl/Mem_Tables.cpp b/plugins/zzogl-pg/opengl/Mem_Tables.cpp
index 620cc72090..f3c67eb8d7 100644
--- a/plugins/zzogl-pg/opengl/Mem_Tables.cpp
+++ b/plugins/zzogl-pg/opengl/Mem_Tables.cpp
@@ -18,6 +18,7 @@
  */
 
 #include "GS.h"
+#include "Mem.h"
 
 u32 g_blockTable32[4][8] =
 {
@@ -247,3 +248,52 @@ u32 g_pageTable16Z[64][64];
 u32 g_pageTable16SZ[64][64];
 u32 g_pageTable8[64][128];
 u32 g_pageTable4[128][128];
+
+/* PSM reference array
+{ 	32, 24, 16, NULL, NULL, NULL, NULL, NULL,
+	NULL, NULL, 16S, NULL, NULL, NULL, NULL, NULL,
+	NULL, NULL, NULL, 8, 4, NULL, NULL, NULL,
+	NULL, NULL, NULL, 8H, NULL, NULL, NULL, NULL,
+	NULL, NULL, NULL, NULL, 4HL, NULL, NULL, NULL,
+	NULL, NULL, NULL, NULL, 4HH, NULL, NULL, NULL,
+	32Z, 24Z, 16Z, NULL, NULL, NULL, NULL, NULL,
+	NULL, NULL, 16SZ, NULL, NULL, NULL, NULL, NULL };
+*/
+
+_getPixelAddress_0 getPixelFun_0[64] = 
+{ 	
+	getPixelAddress32_0, getPixelAddress24_0, getPixelAddress16_0, NULL, NULL, NULL, NULL, NULL,
+	NULL, NULL, getPixelAddress16S_0, NULL, NULL, NULL, NULL, NULL,
+	NULL, NULL, NULL, getPixelAddress8_0, getPixelAddress4_0, NULL, NULL, NULL,
+	NULL, NULL, NULL, getPixelAddress8H_0, NULL, NULL, NULL, NULL,
+	NULL, NULL, NULL, NULL, getPixelAddress4HL_0, NULL, NULL, NULL,
+	NULL, NULL, NULL, NULL, getPixelAddress4HH_0, NULL, NULL, NULL,
+	getPixelAddress32Z_0, getPixelAddress24Z_0, getPixelAddress16Z_0, NULL, NULL, NULL, NULL, NULL,
+	NULL, NULL, getPixelAddress16SZ_0, NULL, NULL, NULL, NULL, NULL 
+};
+
+_writePixel_0 writePixelFun_0[64] = 
+{ 	
+	writePixel32_0, writePixel24_0, writePixel16_0, NULL, NULL, NULL, NULL, NULL,
+	NULL, NULL, writePixel16S_0, NULL, NULL, NULL, NULL, NULL,
+	NULL, NULL, NULL, writePixel8_0, writePixel4_0, NULL, NULL, NULL,
+	NULL, NULL, NULL, writePixel8H_0, NULL, NULL, NULL, NULL,
+	NULL, NULL, NULL, NULL, writePixel4HL_0, NULL, NULL, NULL,
+	NULL, NULL, NULL, NULL, writePixel4HH_0, NULL, NULL, NULL,
+	writePixel32Z_0, writePixel24Z_0, writePixel16Z_0, NULL, NULL, NULL, NULL, NULL,
+	NULL, NULL, writePixel16SZ_0, NULL, NULL, NULL, NULL, NULL 
+};
+
+_readPixel_0 readPixelFun_0[64] = 
+{ 	
+	readPixel32_0, readPixel24_0, readPixel16_0, NULL, NULL, NULL, NULL, NULL,
+	NULL, NULL, readPixel16S_0, NULL, NULL, NULL, NULL, NULL,
+	NULL, NULL, NULL, readPixel8_0, readPixel4_0, NULL, NULL, NULL,
+	NULL, NULL, NULL, readPixel8H_0, NULL, NULL, NULL, NULL,
+	NULL, NULL, NULL, NULL, readPixel4HL_0, NULL, NULL, NULL,
+	NULL, NULL, NULL, NULL, readPixel4HH_0, NULL, NULL, NULL,
+	readPixel32Z_0, readPixel24Z_0, readPixel16Z_0, NULL, NULL, NULL, NULL, NULL,
+	NULL, NULL, readPixel16SZ_0, NULL, NULL, NULL, NULL, NULL 
+};
+
+
diff --git a/plugins/zzogl-pg/opengl/NewRegs.cpp b/plugins/zzogl-pg/opengl/NewRegs.cpp
index bf670b0e31..2646ce066b 100644
--- a/plugins/zzogl-pg/opengl/NewRegs.cpp
+++ b/plugins/zzogl-pg/opengl/NewRegs.cpp
@@ -125,27 +125,16 @@ void __fastcall GIFPackedRegHandlerUV(u32* data)
 	ZZLog::Greg_Log("Packed UV: 0x%x, 0x%x", r->U, r->V);
 }
 
-void __forceinline KICK_VERTEX2()
+void __forceinline KickVertex(bool adc)
 {
 	FUNCLOG
-
-	if (++gs.primC >= (int)g_primmult[prim->prim])
-	{
-		if (NoHighlights(prim->ctxt)) (*ZeroGS::drawfn[prim->prim])();
-
-		gs.primC -= g_primsub[prim->prim];
-	}
-}
-
-void __forceinline KICK_VERTEX3()
-{
-	FUNCLOG
-
 	if (++gs.primC >= (int)g_primmult[prim->prim])
 	{
+		if (!adc && NoHighlights(prim->ctxt)) (*ZeroGS::drawfn[prim->prim])();
+		
 		gs.primC -= g_primsub[prim->prim];
 
-		if (prim->prim == 5)
+		if (adc && prim->prim == 5)
 		{
 			/* tri fans need special processing */
 			if (gs.nTriFanVert == gs.primIndex)
@@ -161,14 +150,7 @@ void __fastcall GIFPackedRegHandlerXYZF2(u32* data)
 	gs.add_vertex(r->X, r->Y,r->Z, r->F);
 
 	// Fix Vertexes up later.
-	if (data[3] & 0x8000)
-	{
-		KICK_VERTEX3();
-	}
-	else
-	{
-		KICK_VERTEX2();
-	}
+	KickVertex(!!(r->ADC));
 	ZZLog::Greg_Log("Packed XYZF2: 0x%x, 0x%x, 0x%x, %f", r->X, r->Y, r->Z, r->F);
 }
 
@@ -179,14 +161,7 @@ void __fastcall GIFPackedRegHandlerXYZ2(u32* data)
 	gs.add_vertex(r->X, r->Y,r->Z);
 
 	// Fix Vertexes up later.
-	if (data[3] & 0x8000)
-	{
-		KICK_VERTEX3();
-	}
-	else
-	{
-		KICK_VERTEX2();
-	}
+	KickVertex(!!(r->ADC));
 	ZZLog::Greg_Log("Packed XYZ2: 0x%x, 0x%x, 0x%x", r->X, r->Y, r->Z);
 }
 
@@ -285,7 +260,7 @@ void __fastcall GIFRegHandlerXYZF2(u32* data)
 	GIFRegXYZF* r = (GIFRegXYZF*)(data);
 	gs.add_vertex(r->X, r->Y,r->Z, r->F);
 
-	KICK_VERTEX2();
+	KickVertex(false);
 	ZZLog::Greg_Log("XYZF2: 0x%x, 0x%x, 0x%x, %f", r->X, r->Y, r->Z, r->F);
 }
 
@@ -295,7 +270,7 @@ void __fastcall GIFRegHandlerXYZ2(u32* data)
 	GIFRegXYZ* r = (GIFRegXYZ*)(data);
 	gs.add_vertex(r->X, r->Y,r->Z);
 
-	KICK_VERTEX2();
+	KickVertex(false);
 	ZZLog::Greg_Log("XYZF2: 0x%x, 0x%x, 0x%x", r->X, r->Y, r->Z);
 }
 
@@ -382,7 +357,7 @@ void __fastcall GIFRegHandlerXYZF3(u32* data)
 	GIFRegXYZF* r = (GIFRegXYZF*)(data);
 	gs.add_vertex(r->X, r->Y,r->Z, r->F);
 
-	KICK_VERTEX3();
+	KickVertex(true);
 	ZZLog::Greg_Log("XYZF3: 0x%x, 0x%x, 0x%x, %f", r->X, r->Y, r->Z, r->F);
 }
 
@@ -392,7 +367,7 @@ void __fastcall GIFRegHandlerXYZ3(u32* data)
 	GIFRegXYZ* r = (GIFRegXYZ*)(data);
 	gs.add_vertex(r->X, r->Y,r->Z);
 
-	KICK_VERTEX3();
+	KickVertex(true);
 	ZZLog::Greg_Log("XYZ3: 0x%x, 0x%x, 0x%x", r->X, r->Y, r->Z);
 }
 
@@ -915,6 +890,7 @@ void __fastcall GIFRegHandlerTRXREG(u32* data)
 void __fastcall GIFRegHandlerTRXDIR(u32* data)
 {
 	FUNCLOG
+	GIFRegTRXDIR* r = (GIFRegTRXDIR*)(data);
 	// Oh dear...
 	
 	// terminate any previous transfers
@@ -934,7 +910,7 @@ void __fastcall GIFRegHandlerTRXDIR(u32* data)
 	gs.dstbuf = gs.dstbufnew;
 	gs.trxpos = gs.trxposnew;
 	
-	gs.imageTransfer = data[0] & 0x3;
+	gs.imageTransfer = r->XDIR;
 	gs.imageWnew = gs.imageWtemp;
 	gs.imageHnew = gs.imageHtemp;
 
diff --git a/plugins/zzogl-pg/opengl/ZZLog.h b/plugins/zzogl-pg/opengl/ZZLog.h
index da299266ed..443e71435b 100644
--- a/plugins/zzogl-pg/opengl/ZZLog.h
+++ b/plugins/zzogl-pg/opengl/ZZLog.h
@@ -1,197 +1,193 @@
-/*  ZZ Open GL graphics plugin
- *  Copyright (c)2009-2010 zeydlitz@gmail.com, arcum42@gmail.com
- *  Based on Zerofrog's ZeroGS KOSMOS (c)2005-2008
- *
- *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
- *
- *  This program is distributed in the hope that it will be useful,
- *  but WITHOUT ANY WARRANTY; without even the implied warranty of
- *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- *  GNU General Public License for more details.
- *
- *  You should have received a copy of the GNU General Public License
- *  along with this program; if not, write to the Free Software
- *  Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA
- */
- 
+/*  ZZ Open GL graphics plugin
+ *  Copyright (c)2009-2010 zeydlitz@gmail.com, arcum42@gmail.com
+ *  Based on Zerofrog's ZeroGS KOSMOS (c)2005-2008
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA
+ */
+ 
 #ifndef ZZLOG_H_INCLUDED
 #define ZZLOG_H_INCLUDED
-
-#include "Util.h"
-
-
-//Logging for errors that are called often should have a time counter.
-#ifdef __LINUX__
-static u32 __attribute__((unused)) lasttime = 0;
-static u32 __attribute__((unused)) BigTime = 5000;
-static bool __attribute__((unused)) SPAM_PASS;
-#else
-static u32 lasttime = 0;
-static u32 BigTime = 5000;
-static bool SPAM_PASS;
-#endif
-
-#define ERROR_LOG_SPAM(text) { \
-	if( timeGetTime() - lasttime > BigTime ) { \
-		ZZLog::Error_Log(text); \
-		lasttime = timeGetTime(); \
-	} \
-}
-// The same macro with one-argument substitution.
-#define ERROR_LOG_SPAMA(fmt, value) { \
-	if( timeGetTime() - lasttime > BigTime ) { \
-		ZZLog::Error_Log(fmt, value); \
-		lasttime = timeGetTime(); \
-	} \
-}
-
-#define ERROR_LOG_SPAM_TEST(text) {\
-	if( timeGetTime() - lasttime > BigTime ) { \
-		ZZLog::Error_Log(text); \
-		lasttime = timeGetTime(); \
-		SPAM_PASS = true; \
-	} \
-	else \
-		SPAM_PASS = false; \
-}
-
-#if DEBUG_PROF
-#define FILE_IS_IN_CHECK ((strcmp(__FILE__, "targets.cpp") == 0) || (strcmp(__FILE__, "ZZoglFlush.cpp") == 0))
-
-#define FUNCLOG {\
-	static bool Was_Here = false; \
-	static unsigned long int waslasttime = 0; \
-	if (!Was_Here && FILE_IS_IN_CHECK) { \
-		Was_Here = true;\
-		ZZLog::Error_Log("%s:%d %s", __FILE__, __LINE__, __func__); \
-		waslasttime = timeGetTime(); \
-	} \
-	if (FILE_IS_IN_CHECK && (timeGetTime() - waslasttime > BigTime ))  { \
-		Was_Here = false; \
-	} \
-}
-#else
-#define FUNCLOG
-#endif
-
-//#define WRITE_GREG_LOGS
-//#define WRITE_PRIM_LOGS
-#if defined(_DEBUG) && !defined(ZEROGS_DEVBUILD)
-#define ZEROGS_DEVBUILD
-#endif
-
-#ifdef ZEROGS_DEVBUILD
-//#define DEVBUILD
-#endif
-
-
-// sends a message to output window if assert fails
-#define BMSG(x, str)			{ if( !(x) ) { ZZLog::Log(str); ZZLog::Log(str); } }
-#define BMSG_RETURN(x, str)	{ if( !(x) ) { ZZLog::Log(str); ZZLog::Log(str); return; } }
-#define BMSG_RETURNX(x, str, rtype)	{ if( !(x) ) { ZZLog::Log(str); ZZLog::Log(str); return (##rtype); } }
-#define B(x)				{ if( !(x) ) { ZZLog::Log(_#x"\n"); ZZLog::Log(#x"\n"); } }
-#define B_RETURN(x)			{ if( !(x) ) { ZZLog::Error_Log("%s:%d: %s", __FILE__, (u32)__LINE__, #x); return; } }
-#define B_RETURNX(x, rtype)			{ if( !(x) ) { ZZLog::Error_Log("%s:%d: %s", __FILE__, (u32)__LINE__, #x); return (##rtype); } }
-#define B_G(x, action)			{ if( !(x) ) { ZZLog::Error_Log("%s:%d: %s", __FILE__, (u32)__LINE__, #x); action; } }
-
-#define GL_REPORT_ERROR() \
-{ \
-	GLenum err = glGetError(); \
-	if( err != GL_NO_ERROR ) \
-	{ \
-		ZZLog::Error_Log("%s:%d: gl error %s(0x%x)", __FILE__, (int)__LINE__, error_name(err), err); \
-		ZeroGS::HandleGLError(); \
-	} \
-}
-
-#ifdef _DEBUG
-#	define GL_REPORT_ERRORD() \
-{ \
-	GLenum err = glGetError(); \
-	if( err != GL_NO_ERROR ) \
-	{ \
-		ZZLog::Error_Log("%s:%d: gl error %s (0x%x)", __FILE__, (int)__LINE__, error_name(err), err); \
-		ZeroGS::HandleGLError(); \
-	} \
-}
-#else
-#	define GL_REPORT_ERRORD()
-#endif
-
-
-inline const char *error_name(int err)
-{
-	switch (err)
-	{
-		case GL_NO_ERROR:
-			return "GL_NO_ERROR";
-
-		case GL_INVALID_ENUM:
-			return "GL_INVALID_ENUM";
-
-		case GL_INVALID_VALUE:
-			return "GL_INVALID_VALUE";
-
-		case GL_INVALID_OPERATION:
-			return "GL_INVALID_OPERATION";
-
-		case GL_STACK_OVERFLOW:
-			return "GL_STACK_OVERFLOW";
-
-		case GL_STACK_UNDERFLOW:
-			return "GL_STACK_UNDERFLOW";
-
-		case GL_OUT_OF_MEMORY:
-			return "GL_OUT_OF_MEMORY";
-
-		case GL_TABLE_TOO_LARGE:
-			return "GL_TABLE_TOO_LARGE";
-
-		case GL_INVALID_FRAMEBUFFER_OPERATION:
-			return "GL_INVALID_FRAMEBUFFER_OPERATION";
-			
-		default:
-			return "Unknown GL error";
-	}
-}
-
-extern void __LogToConsole(const char *fmt, ...);
-
-// Subset of zerogs, to avoid that whole huge header.
-namespace ZeroGS
-{
-extern void AddMessage(const char* pstr, u32 ms);
-extern void SetAA(int mode);
-extern void SetNegAA(int mode);
-extern bool Create(int width, int height);
-extern void Destroy(bool bD3D);
-extern void StartCapture();
-extern void StopCapture();
-}
-
-namespace ZZLog
-{
-extern bool IsLogging();
-void SetDir(const char* dir);
-extern bool Open();
-extern void Close();
-extern void Message(const char *fmt, ...);
-extern void Log(const char *fmt, ...);
-void WriteToScreen(const char* pstr, u32 ms = 5000);
-extern void WriteToConsole(const char *fmt, ...);
-extern void Print(const char *fmt, ...);
-extern void WriteLn(const char *fmt, ...);
-
-extern void Greg_Log(const char *fmt, ...);
-extern void Prim_Log(const char *fmt, ...);
-extern void GS_Log(const char *fmt, ...);
-
-extern void Debug_Log(const char *fmt, ...);
-extern void Warn_Log(const char *fmt, ...);
-extern void Error_Log(const char *fmt, ...);
+
+#include "Util.h"
+
+
+//Logging for errors that are called often should have a time counter.
+#ifdef __LINUX__
+static u32 __attribute__((unused)) lasttime = 0;
+static u32 __attribute__((unused)) BigTime = 5000;
+static bool __attribute__((unused)) SPAM_PASS;
+#else
+static u32 lasttime = 0;
+static u32 BigTime = 5000;
+static bool SPAM_PASS;
+#endif
+
+#define ERROR_LOG_SPAM(text) { \
+	if( timeGetTime() - lasttime > BigTime ) { \
+		ZZLog::Error_Log(text); \
+		lasttime = timeGetTime(); \
+	} \
+}
+// The same macro with one-argument substitution.
+#define ERROR_LOG_SPAMA(fmt, value) { \
+	if( timeGetTime() - lasttime > BigTime ) { \
+		ZZLog::Error_Log(fmt, value); \
+		lasttime = timeGetTime(); \
+	} \
+}
+
+#define ERROR_LOG_SPAM_TEST(text) {\
+	if( timeGetTime() - lasttime > BigTime ) { \
+		ZZLog::Error_Log(text); \
+		lasttime = timeGetTime(); \
+		SPAM_PASS = true; \
+	} \
+	else \
+		SPAM_PASS = false; \
+}
+
+#if DEBUG_PROF
+#define FILE_IS_IN_CHECK ((strcmp(__FILE__, "targets.cpp") == 0) || (strcmp(__FILE__, "ZZoglFlush.cpp") == 0))
+
+#define FUNCLOG {\
+	static bool Was_Here = false; \
+	static unsigned long int waslasttime = 0; \
+	if (!Was_Here && FILE_IS_IN_CHECK) { \
+		Was_Here = true;\
+		ZZLog::Error_Log("%s:%d %s", __FILE__, __LINE__, __func__); \
+		waslasttime = timeGetTime(); \
+	} \
+	if (FILE_IS_IN_CHECK && (timeGetTime() - waslasttime > BigTime ))  { \
+		Was_Here = false; \
+	} \
+}
+#else
+#define FUNCLOG
+#endif
+
+//#define WRITE_GREG_LOGS
+//#define WRITE_PRIM_LOGS
+#if defined(_DEBUG) && !defined(ZEROGS_DEVBUILD)
+#define ZEROGS_DEVBUILD
+#endif
+
+
+// sends a message to output window if assert fails
+#define BMSG(x, str)			{ if( !(x) ) { ZZLog::Log(str); ZZLog::Log(str); } }
+#define BMSG_RETURN(x, str)	{ if( !(x) ) { ZZLog::Log(str); ZZLog::Log(str); return; } }
+#define BMSG_RETURNX(x, str, rtype)	{ if( !(x) ) { ZZLog::Log(str); ZZLog::Log(str); return (##rtype); } }
+#define B(x)				{ if( !(x) ) { ZZLog::Log(_#x"\n"); ZZLog::Log(#x"\n"); } }
+#define B_RETURN(x)			{ if( !(x) ) { ZZLog::Error_Log("%s:%d: %s", __FILE__, (u32)__LINE__, #x); return; } }
+#define B_RETURNX(x, rtype)			{ if( !(x) ) { ZZLog::Error_Log("%s:%d: %s", __FILE__, (u32)__LINE__, #x); return (##rtype); } }
+#define B_G(x, action)			{ if( !(x) ) { ZZLog::Error_Log("%s:%d: %s", __FILE__, (u32)__LINE__, #x); action; } }
+
+#define GL_REPORT_ERROR() \
+{ \
+	GLenum err = glGetError(); \
+	if( err != GL_NO_ERROR ) \
+	{ \
+		ZZLog::Error_Log("%s:%d: gl error %s(0x%x)", __FILE__, (int)__LINE__, error_name(err), err); \
+		ZeroGS::HandleGLError(); \
+	} \
+}
+
+#ifdef _DEBUG
+#	define GL_REPORT_ERRORD() \
+{ \
+	GLenum err = glGetError(); \
+	if( err != GL_NO_ERROR ) \
+	{ \
+		ZZLog::Error_Log("%s:%d: gl error %s (0x%x)", __FILE__, (int)__LINE__, error_name(err), err); \
+		ZeroGS::HandleGLError(); \
+	} \
+}
+#else
+#	define GL_REPORT_ERRORD()
+#endif
+
+
+inline const char *error_name(int err)
+{
+	switch (err)
+	{
+		case GL_NO_ERROR:
+			return "GL_NO_ERROR";
+
+		case GL_INVALID_ENUM:
+			return "GL_INVALID_ENUM";
+
+		case GL_INVALID_VALUE:
+			return "GL_INVALID_VALUE";
+
+		case GL_INVALID_OPERATION:
+			return "GL_INVALID_OPERATION";
+
+		case GL_STACK_OVERFLOW:
+			return "GL_STACK_OVERFLOW";
+
+		case GL_STACK_UNDERFLOW:
+			return "GL_STACK_UNDERFLOW";
+
+		case GL_OUT_OF_MEMORY:
+			return "GL_OUT_OF_MEMORY";
+
+		case GL_TABLE_TOO_LARGE:
+			return "GL_TABLE_TOO_LARGE";
+
+		case GL_INVALID_FRAMEBUFFER_OPERATION:
+			return "GL_INVALID_FRAMEBUFFER_OPERATION";
+			
+		default:
+			return "Unknown GL error";
+	}
+}
+
+extern void __LogToConsole(const char *fmt, ...);
+
+// Subset of zerogs, to avoid that whole huge header.
+namespace ZeroGS
+{
+extern void AddMessage(const char* pstr, u32 ms);
+extern void SetAA(int mode);
+extern void SetNegAA(int mode);
+extern bool Create(int width, int height);
+extern void Destroy(bool bD3D);
+extern void StartCapture();
+extern void StopCapture();
+}
+
+namespace ZZLog
+{
+extern bool IsLogging();
+void SetDir(const char* dir);
+extern bool Open();
+extern void Close();
+extern void Message(const char *fmt, ...);
+extern void Log(const char *fmt, ...);
+void WriteToScreen(const char* pstr, u32 ms = 5000);
+extern void WriteToConsole(const char *fmt, ...);
+extern void Print(const char *fmt, ...);
+extern void WriteLn(const char *fmt, ...);
+
+extern void Greg_Log(const char *fmt, ...);
+extern void Prim_Log(const char *fmt, ...);
+extern void GS_Log(const char *fmt, ...);
+
+extern void Debug_Log(const char *fmt, ...);
+extern void Warn_Log(const char *fmt, ...);
+extern void Error_Log(const char *fmt, ...);
 };
 
 #endif // ZZLOG_H_INCLUDED
diff --git a/plugins/zzogl-pg/opengl/ZZoglFlush.cpp b/plugins/zzogl-pg/opengl/ZZoglFlush.cpp
index 29ea9546e2..e4338b555a 100644
--- a/plugins/zzogl-pg/opengl/ZZoglFlush.cpp
+++ b/plugins/zzogl-pg/opengl/ZZoglFlush.cpp
@@ -29,7 +29,7 @@
 using namespace ZeroGS;
 
 //------------------ Defines
-#ifndef DEVBUILD
+#ifndef ZEROGS_DEVBUILD
 
 #define INC_GENVARS()
 #define INC_TEXVARS()
diff --git a/plugins/zzogl-pg/opengl/targets.cpp b/plugins/zzogl-pg/opengl/targets.cpp
index 6adf5e690f..41e0ccbabf 100644
--- a/plugins/zzogl-pg/opengl/targets.cpp
+++ b/plugins/zzogl-pg/opengl/targets.cpp
@@ -42,10 +42,6 @@ extern bool g_bUpdateStencil;
 #	define INC_RESOLVE() ++g_nResolve
 #endif
 
-#ifdef DEVBUILD
-//static int g_bSaveResolved = 0;
-#endif
-
 extern int s_nResolved;
 extern u32 g_nResolve;
 extern bool g_bSaveTrans;
@@ -250,6 +246,7 @@ void ZeroGS::CRenderTarget::SetTarget(int fbplocal, const Rect2& scissor, int co
 	}
 
 	// set render states
+	// Bleh. I *really* need to fix this. << 3 when setting the scissors, then >> 3 when using them... --Arcum42
 	scissorrect.x = scissor.x0 >> 3;
 	scissorrect.y = (scissor.y0 >> 3) + dy;
 	scissorrect.w = (scissor.x1 >> 3) + 1;
@@ -302,7 +299,7 @@ void ZeroGS::CRenderTarget::Resolve()
 
 		GL_REPORT_ERRORD();
 
-#if defined(DEVBUILD)
+#if defined(ZEROGS_DEVBUILD)
 
 		if (g_bSaveResolved)
 		{
@@ -328,7 +325,7 @@ void ZeroGS::CRenderTarget::Resolve(int startrange, int endrange)
 		// flush if necessary
 		FlushIfNecesary(this) ;
 
-#if defined(DEVBUILD)
+#if defined(ZEROGS_DEVBUILD)
 		if (g_bSaveResolved)
 		{
 			SaveTexture("resolved.tga", GL_TEXTURE_RECTANGLE_NV, ptex, RW(fbw), RH(fbh));
@@ -2290,10 +2287,12 @@ ZeroGS::CMemoryTarget* ZeroGS::CMemoryTargetMngr::GetMemoryTarget(const tex0Info
 			{
 				// This is not unusual situation, when vector<u8> does not 16bit alignment, that is destructive for SSE2
 				// instruction movdqa [%eax], xmm0
-				// The idea would be resise vector to 15 elements, that set ptxedata to aligned position.
+				// The idea would be resize vector to 15 elements, that set ptxedata to aligned position.
 				// Later we would move eax by 16, so only we should verify is first element align
 				// FIXME. As I see, texdata used only once here, it does not have any impact on other code.
 				// Probably, usage of _aligned_maloc() would be preferable.
+				
+				// Note: this often happens when changing AA.
 				int disalignment = 16 - ((u32)(uptr)dst) % 16;		// This is value of shift. It could be 0 < disalignment <= 15
 				ptexdata = &texdata[disalignment];			// Set pointer to aligned element
 				dst = (u16*)ptexdata;
diff --git a/plugins/zzogl-pg/opengl/x86.cpp b/plugins/zzogl-pg/opengl/x86.cpp
index 581d023139..d32718dbfb 100644
--- a/plugins/zzogl-pg/opengl/x86.cpp
+++ b/plugins/zzogl-pg/opengl/x86.cpp
@@ -308,13 +308,10 @@ extern "C" void __fastcall WriteCLUT_T32_I4_CSM1_sse2(u32* vm, u32* clut)
 }
 
 
-extern "C"
-{
-	PCSX2_ALIGNED16(int s_clut16mask2[4]) = { 0x0000ffff, 0x0000ffff, 0x0000ffff, 0x0000ffff };
-	PCSX2_ALIGNED16(int s_clut16mask[8]) = { 0xffff0000, 0xffff0000, 0xffff0000, 0xffff0000,
+static const __aligned16 int s_clut16mask2[4] = { 0x0000ffff, 0x0000ffff, 0x0000ffff, 0x0000ffff };
+static const __aligned16 int s_clut16mask[8] = { 0xffff0000, 0xffff0000, 0xffff0000, 0xffff0000,
 										   0x0000ffff, 0x0000ffff, 0x0000ffff, 0x0000ffff
 										   };
-}
 
 extern "C" void __fastcall WriteCLUT_T16_I4_CSM1_sse2(u32* vm, u32* clut)
 {
@@ -433,11 +430,11 @@ WriteUnaligned:
 End:
 	}
 #else
-	__asm__(".intel_syntax noprefix\n"
-	"movdqa xmm0, xmmword ptr [ecx]\n"
-	"movdqa xmm1, xmmword ptr [ecx+16]\n"
-	"movdqa xmm2, xmmword ptr [ecx+32]\n"
-	"movdqa xmm3, xmmword ptr [ecx+48]\n"
+	__asm__ __volatile__(".intel_syntax noprefix\n"
+	"movdqa xmm0, xmmword ptr [%[vm]]\n"
+	"movdqa xmm1, xmmword ptr [%[vm]+16]\n"
+	"movdqa xmm2, xmmword ptr [%[vm]+32]\n"
+	"movdqa xmm3, xmmword ptr [%[vm]+48]\n"
 
 	// rearrange
 	"pshuflw xmm0, xmm0, 0x88\n"
@@ -457,14 +454,14 @@ End:
 
 	"pxor xmm6, xmm6\n"
 
-	"test edx, 15\n"
+	"test %[clut], 15\n"
 	"jnz WriteUnaligned\n"
 
-	"movdqa xmm7, [s_clut16mask]\n" // saves upper 16 bits
+	"movdqa xmm7, %[s_clut16mask]\n" // saves upper 16 bits
 
 	// have to save interlaced with the old data
-	"movdqa xmm4, [edx]\n"
-	"movdqa xmm5, [edx+32]\n"
+	"movdqa xmm4, [%[clut]]\n"
+	"movdqa xmm5, [%[clut]+32]\n"
 	"movhlps xmm1, xmm0\n"
 	"movlhps xmm0, xmm2\n"// lower 8 colors
 
@@ -483,29 +480,29 @@ End:
 	"punpckhwd xmm2, xmm6\n"
 	"punpckhwd xmm3, xmm6\n"
 
-	"movdqa [edx], xmm0\n"
-	"movdqa [edx+32], xmm1\n"
+	"movdqa [%[clut]], xmm0\n"
+	"movdqa [%[clut]+32], xmm1\n"
 
 	"movdqa xmm5, xmm7\n"
-	"pand xmm7, [edx+16]\n"
-	"pand xmm5, [edx+48]\n"
+	"pand xmm7, [%[clut]+16]\n"
+	"pand xmm5, [%[clut]+48]\n"
 
 	"por xmm2, xmm7\n"
 	"por xmm3, xmm5\n"
 
-	"movdqa [edx+16], xmm2\n"
-	"movdqa [edx+48], xmm3\n"
+	"movdqa [%[clut]+16], xmm2\n"
+	"movdqa [%[clut]+48], xmm3\n"
 	"jmp WriteCLUT_T16_I4_CSM1_End\n"
 
 	"WriteUnaligned:\n"
-	// %edx is offset by 2
-	"sub edx, 2\n"
+	// %[clut] is offset by 2
+	"sub %[clut], 2\n"
 
-	"movdqa xmm7, [[s_clut16mask2]]\n" // saves lower 16 bits
+	"movdqa xmm7, %[s_clut16mask2]\n" // saves lower 16 bits
 
 	// have to save interlaced with the old data
-	"movdqa xmm4, [edx]\n"
-	"movdqa xmm5, [edx+32]\n"
+	"movdqa xmm4, [%[clut]]\n"
+	"movdqa xmm5, [%[clut]+32]\n"
 	"movhlps xmm1, xmm0\n"
 	"movlhps xmm0, xmm2\n" // lower 8 colors
 
@@ -528,24 +525,24 @@ End:
 	"pslld xmm2, 16\n"
 	"pslld xmm3, 16\n"
 
-	"movdqa [edx], xmm0\n"
-	"movdqa [edx+32], xmm1\n"
+	"movdqa [%[clut]], xmm0\n"
+	"movdqa [%[clut]+32], xmm1\n"
 
 	"movdqa xmm5, xmm7\n"
-	"pand xmm7, [edx+16]\n"
-	"pand xmm5, [edx+48]\n"
+	"pand xmm7, [%[clut]+16]\n"
+	"pand xmm5, [%[clut]+48]\n"
 
 	"por xmm2, xmm7\n"
 	"por xmm3, xmm5\n"
 
-	"movdqa [edx+16], xmm2\n"
-	"movdqa [edx+48], xmm3\n"
+	"movdqa [%[clut]+16], xmm2\n"
+	"movdqa [%[clut]+48], xmm3\n"
 	"WriteCLUT_T16_I4_CSM1_End:\n"
 	"\n"
 	".att_syntax\n" 
-    : [s_clut16mask] "=m" (s_clut16mask), [s_clut16mask2] "=m" (s_clut16mask2)
-    : "c" (vm), "d" (clut)
-    : "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7"
+    :
+    : [vm] "r" (vm), [clut] "r" (clut), [s_clut16mask] "m" (*s_clut16mask), [s_clut16mask2] "m" (*s_clut16mask2)
+    : "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7", "memory"
        );
 #endif // _MSC_VER
 }
@@ -718,15 +715,15 @@ Z16Loop:
 	}
 #else // _MSC_VER
 
-	__asm__(".intel_syntax\n"
+	__asm__ __volatile__(".intel_syntax\n"
 	"pxor %%xmm7, %%xmm7\n"
 
 	"Z16Loop:\n"
 	// unpack 64 bytes at a time
-	"movdqa %%xmm0, [%0]\n"
-	"movdqa %%xmm2, [%0+16]\n"
-	"movdqa %%xmm4, [%0+32]\n"
-	"movdqa %%xmm6, [%0+48]\n"
+	"movdqa %%xmm0, [%[src]]\n"
+	"movdqa %%xmm2, [%[src]+16]\n"
+	"movdqa %%xmm4, [%[src]+32]\n"
+	"movdqa %%xmm6, [%[src]+48]\n"
 
 	"movdqa %%xmm1, %%xmm0\n"
 	"movdqa %%xmm3, %%xmm2\n"
@@ -738,35 +735,35 @@ Z16Loop:
 	"punpckhwd %%xmm3, %%xmm7\n"
 
 	// start saving
-	"movdqa [%1], %%xmm0\n"
-	"movdqa [%1+16], %%xmm1\n"
+	"movdqa [%[dst]], %%xmm0\n"
+	"movdqa [%[dst]+16], %%xmm1\n"
 
 	"punpcklwd %%xmm4, %%xmm7\n"
 	"punpckhwd %%xmm5, %%xmm7\n"
 
-	"movdqa [%1+32], %%xmm2\n"
-	"movdqa [%1+48], %%xmm3\n"
+	"movdqa [%[dst]+32], %%xmm2\n"
+	"movdqa [%[dst]+48], %%xmm3\n"
 
 	"movdqa %%xmm0, %%xmm6\n"
 	"punpcklwd %%xmm6, %%xmm7\n"
 
-	"movdqa [%1+64], %%xmm4\n"
-	"movdqa [%1+80], %%xmm5\n"
+	"movdqa [%[dst]+64], %%xmm4\n"
+	"movdqa [%[dst]+80], %%xmm5\n"
 
 	"punpckhwd %%xmm0, %%xmm7\n"
 
-	"movdqa [%1+96], %%xmm6\n"
-	"movdqa [%1+112], %%xmm0\n"
+	"movdqa [%[dst]+96], %%xmm6\n"
+	"movdqa [%[dst]+112], %%xmm0\n"
 
-	"add %0, 64\n"
-	"add %1, 128\n"
-	"sub %2, 1\n"
+	"add %[src], 64\n"
+	"add %[dst], 128\n"
+	"sub %[iters], 1\n"
 	"jne Z16Loop\n"
 
 ".att_syntax\n" 
-    : "=r"(src), "=r"(dst), "=r"(iters) 
-    : "0"(src), "1"(dst), "2"(iters)
-    : "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7"
+    : "=&r"(src), "=&r"(dst), "=&r"(iters)
+    : [src] "0"(src), [dst] "1"(dst), [iters] "2"(iters)
+    : "xmm0", "xmm1", "xmm2", "xmm3", "xmm4", "xmm5", "xmm6", "xmm7", "memory"
        );
 #endif // _MSC_VER
 }
diff --git a/plugins/zzogl-pg/opengl/zerogs.cpp b/plugins/zzogl-pg/opengl/zerogs.cpp
index 2d04ed3e27..c311e7771b 100644
--- a/plugins/zzogl-pg/opengl/zerogs.cpp
+++ b/plugins/zzogl-pg/opengl/zerogs.cpp
@@ -803,15 +803,10 @@ void ZeroGS::KickSprite()
 
 	int next = (gs.primIndex + 1) % ARRAY_SIZE(gs.gsvertex);
 	int last = (gs.primIndex + 2) % ARRAY_SIZE(gs.gsvertex);
-
-	// sprite is too small and AA shows lines (tek4)
-
-	if (s_AAx)
-	{
-		gs.gsvertex[last].x += 4;
-
-		if (s_AAy) gs.gsvertex[last].y += 4;
-	}
+	
+	// sprite is too small and AA shows lines (tek4, Mana Khemia)
+	gs.gsvertex[last].x += (4*s_AAx);
+	gs.gsvertex[last].y += (4*s_AAy);
 
 	// might be bad sprite (KH dialog text)
 	//if( gs.gsvertex[next].x == gs.gsvertex[last].x || gs.gsvertex[next].y == gs.gsvertex[last].y )
diff --git a/plugins/zzogl-pg/opengl/zerogs.h b/plugins/zzogl-pg/opengl/zerogs.h
index 3c06df3cd5..6491297e3f 100644
--- a/plugins/zzogl-pg/opengl/zerogs.h
+++ b/plugins/zzogl-pg/opengl/zerogs.h
@@ -96,7 +96,7 @@ extern u32 ptexBilinearBlocks;
 // State parameters
 
 
-#ifdef DEVBUILD
+#ifdef ZEROGS_DEVBUILD
 extern char* EFFECT_NAME;
 extern char* EFFECT_DIR;
 extern u32 g_nGenVars, g_nTexVars, g_nAlphaVars, g_nResolve;