diff --git a/plugins/GSdx/GSRendererSW.cpp b/plugins/GSdx/GSRendererSW.cpp
index eca864ac19..dab27400bd 100644
--- a/plugins/GSdx/GSRendererSW.cpp
+++ b/plugins/GSdx/GSRendererSW.cpp
@@ -250,10 +250,12 @@ void GSRendererSW::Draw()
 		GSVertexSW* RESTRICT d = sd->vertex;
 
 		GSVector4i o = (GSVector4i)m_context->XYOFFSET;
-		GSVector4 tsize = GSVector4(0x10000 << m_context->TEX0.TW, 0x10000 << m_context->TEX0.TH);
+		GSVector4 tsize = GSVector4(0x10000 << m_context->TEX0.TW, 0x10000 << m_context->TEX0.TH, 1, 0);
 
 		for(size_t i = 0; i < m_vertex.next; i++, s++, d++)
 		{
+			// TODO: load xyzuvf in one piece
+
 			uint32 z = s->XYZ.Z;
 
 			GSVector4i xy = GSVector4i::load((int)s->XYZ.u32[0]).upl16() - o;
@@ -263,6 +265,8 @@ void GSRendererSW::Draw()
 
 			p = GSVector4(xy).xyxy(GSVector4(zf) + (GSVector4::m_x4f800000 & GSVector4::cast(zf.sra32(31)))) * g_pos_scale;
 
+			GSVector4 stcq = GSVector4::load<true>(&s->m[0]); // s t rgba q
+
 			if(PRIM->TME)
 			{
 				if(PRIM->FST)
@@ -271,12 +275,11 @@ void GSRendererSW::Draw()
 				}
 				else
 				{
-					t = GSVector4(s->ST.S, s->ST.T) * tsize;
-					t = t.xyxy(GSVector4::load(s->RGBAQ.Q));
+					t = stcq.xyww() * tsize;
 				}
 			}
 
-			c = GSVector4::rgba32(s->RGBAQ.u32[0], 7);
+			c = GSVector4(GSVector4i::cast(stcq).zzzz().u8to32() << 7);
 
 			d->p = p;
 			d->c = c;
@@ -284,7 +287,7 @@ void GSRendererSW::Draw()
 
 			if(sd->primclass == GS_SPRITE_CLASS)
 			{
-				d->t.u32[3] = z;
+				d->t.u32[3] = z; // TODO: store this to the 4th unused GSVector4?
 			}
 		}
 	}
diff --git a/plugins/GSdx/GSVector.cpp b/plugins/GSdx/GSVector.cpp
index 7a824cb9ad..27b6a5a0e9 100644
--- a/plugins/GSdx/GSVector.cpp
+++ b/plugins/GSdx/GSVector.cpp
@@ -22,8 +22,9 @@
 #include "stdafx.h"
 #include "GSVector.h"
 
-const GSVector4i GSVector4i::m_xff[16] = 
+const GSVector4i GSVector4i::m_xff[17] = 
 {
+	GSVector4i(0x00000000, 0x00000000, 0x00000000, 0x00000000), 
 	GSVector4i(0x000000ff, 0x00000000, 0x00000000, 0x00000000), 
 	GSVector4i(0x0000ffff, 0x00000000, 0x00000000, 0x00000000), 
 	GSVector4i(0x00ffffff, 0x00000000, 0x00000000, 0x00000000), 
@@ -42,8 +43,9 @@ const GSVector4i GSVector4i::m_xff[16] =
 	GSVector4i(0xffffffff, 0xffffffff, 0xffffffff, 0xffffffff),
 };
 
-const GSVector4i GSVector4i::m_x0f[16] =
+const GSVector4i GSVector4i::m_x0f[17] =
 {
+	GSVector4i(0x00000000, 0x00000000, 0x00000000, 0x00000000), 
 	GSVector4i(0x0000000f, 0x00000000, 0x00000000, 0x00000000), 
 	GSVector4i(0x00000f0f, 0x00000000, 0x00000000, 0x00000000), 
 	GSVector4i(0x000f0f0f, 0x00000000, 0x00000000, 0x00000000), 
diff --git a/plugins/GSdx/GSVector.h b/plugins/GSdx/GSVector.h
index f066a11016..5646dfbd6f 100644
--- a/plugins/GSdx/GSVector.h
+++ b/plugins/GSdx/GSVector.h
@@ -79,8 +79,8 @@ class GSVector4;
 
 __aligned(class, 16) GSVector4i
 {
-	static const GSVector4i m_xff[16];
-	static const GSVector4i m_x0f[16];
+	static const GSVector4i m_xff[17];
+	static const GSVector4i m_x0f[17];
 
 public:
 	union
diff --git a/plugins/GSdx/GSVertexTrace.cpp b/plugins/GSdx/GSVertexTrace.cpp
index bfb0857c48..63f4c1a3bd 100644
--- a/plugins/GSdx/GSVertexTrace.cpp
+++ b/plugins/GSdx/GSVertexTrace.cpp
@@ -140,6 +140,8 @@ void GSVertexTrace::FindMinMax(const void* vertex, const uint32* index, int coun
 	GSVector4i cmin = GSVector4i::xffffffff();
 	GSVector4i cmax = GSVector4i::zero();
 
+	// TODO: SSE41 has integer min/max, use that for xy/z/uv/f
+
 	const GSVertex* RESTRICT v = (GSVertex*)vertex;
 
 	for(int i = 0; i < count; i += n)