Vertex Loader: SSE implementations of more position/texcoord/normal formats

~35-45% faster NFS:HP2, possibly other vertex-bound games.

1 files changed, 7 insertions, 8 deletions

diff --git a/Source/Core/VideoCommon/VertexLoader.cpp b/Source/Core/VideoCommon/VertexLoader.cpp
index 373dfae16b..80a891ff26 100644
--- a/Source/Core/VideoCommon/VertexLoader.cpp
+++ b/Source/Core/VideoCommon/VertexLoader.cpp
@@ -41,8 +41,9 @@ static int s_texmtxread = 0;
 int tcIndex;
 int colIndex;
 int colElements[2];
-float posScale;
-float tcScale[8];
+// Duplicated (4x and 2x respectively) and used in SSE code in the vertex loader JIT
+GC_ALIGNED128(float posScale[4]);
+GC_ALIGNED64(float tcScale[8][2]);
 
 static const float fractionTable[32] = {
 	1.0f / (1U << 0), 1.0f / (1U << 1), 1.0f / (1U << 2), 1.0f / (1U << 3),
@@ -65,10 +66,8 @@ static void LOADERDECL PosMtx_ReadDirect_UByte()
 
 static void LOADERDECL PosMtx_Write()
 {
-	DataWrite<u8>(s_curposmtx);
-	DataWrite<u8>(0);
-	DataWrite<u8>(0);
-	DataWrite<u8>(0);
+	// u8, 0, 0, 0
+	DataWrite<u32>(s_curposmtx);
 }
 
 static void LOADERDECL TexMtx_ReadDirect_UByte()
@@ -451,10 +450,10 @@ void VertexLoader::SetupRunVertices(const VAT& vat, int primitive, int const cou
 	m_VtxAttr.texCoord[6].Frac = vat.g2.Tex6Frac;
 	m_VtxAttr.texCoord[7].Frac = vat.g2.Tex7Frac;
 
-	posScale = fractionTable[m_VtxAttr.PosFrac];
+	posScale[0] = posScale[1] = posScale[2] = posScale[3] = fractionTable[m_VtxAttr.PosFrac];
 	if (m_native_components & VB_HAS_UVALL)
 		for (int i = 0; i < 8; i++)
-			tcScale[i] = fractionTable[m_VtxAttr.texCoord[i].Frac];
+			tcScale[i][0] = tcScale[i][1] = fractionTable[m_VtxAttr.texCoord[i].Frac];
 	for (int i = 0; i < 2; i++)
 		colElements[i] = m_VtxAttr.color[i].Elements;