Added shuffle() variant to the standard library that takes two

varying values and a permutation index that spans the concatenation
of the two of them (along the lines of SHUFPS...)

tests/shuffle2-1.ispc

@@ -0,0 +1,13 @@
+
+export uniform int width() { return programCount; }
+
+export void f_fu(uniform float RET[], uniform float aFOO[], uniform float b) {
+    int32 aa = aFOO[programIndex]; 
+    int32 bb = aa + programCount;
+    int32 shuf = shuffle(aa, bb, programCount + 1);
+    RET[programIndex] = shuf;
+}
+
+export void result(uniform float RET[]) {
+    RET[programIndex] = 2 + programCount;
+}

tests/shuffle2-2.ispc

@@ -0,0 +1,13 @@
+
+export uniform int width() { return programCount; }
+
+export void f_fu(uniform float RET[], uniform float aFOO[], uniform float b) {
+    int32 aa = aFOO[programIndex]; 
+    int32 bb = aa + programCount;
+    int32 shuf = shuffle(aa, bb, programIndex + 2);
+    RET[programIndex] = shuf;
+}
+
+export void result(uniform float RET[]) {
+    RET[programIndex] = 3 + programIndex;
+}

tests/shuffle2-3.ispc

@@ -0,0 +1,13 @@
+
+export uniform int width() { return programCount; }
+
+export void f_fu(uniform float RET[], uniform float aFOO[], uniform float b) {
+    float aa = aFOO[programIndex]; 
+    float bb = aa + programCount;
+    float shuf = shuffle(aa, bb, programIndex + 2 + (int)b - 5);
+    RET[programIndex] = shuf;
+}
+
+export void result(uniform float RET[]) {
+    RET[programIndex] = 3 + programIndex;
+}

tests/shuffle2-4.ispc

@@ -0,0 +1,13 @@
+
+export uniform int width() { return programCount; }
+
+export void f_fu(uniform float RET[], uniform float aFOO[], uniform float b) {
+    double aa = aFOO[programIndex]; 
+    double bb = aa + programCount;
+    double shuf = shuffle(aa, bb, 2*programIndex+(int)b-5);
+    RET[programIndex] = shuf;
+}
+
+export void result(uniform float RET[]) {
+    RET[programIndex] = 1 + 2*programIndex;
+}

tests/shuffle2-5.ispc

@@ -0,0 +1,13 @@
+
+export uniform int width() { return programCount; }
+
+export void f_fu(uniform float RET[], uniform float aFOO[], uniform float b) {
+    double aa = aFOO[programIndex]; 
+    double bb = aa + programCount;
+    double shuf = shuffle(aa, bb, 2*programIndex);
+    RET[programIndex] = shuf;
+}
+
+export void result(uniform float RET[]) {
+    RET[programIndex] = 1 + 2*programIndex;
+}