Tiny cleanups, doc update re int8/16 performance

docs/ispc.txt

@@ -93,6 +93,7 @@ Contents:
   + `Understanding How to Interoperate With the Application's Data`_
   + `Communicating Between SPMD Program Instances`_
   + `Gather and Scatter`_
+  + `8 and 16-bit Integer Types`_
   + `Low-level Vector Tricks`_
   + `Debugging`_
   + `The "Fast math" Option`_
@@ -2528,6 +2529,15 @@ do a vector load.  For example, given:
 
 A regular vector load is done from array, starting at offset ``2*x``.
 
+
+8 and 16-bit Integer Types
+--------------------------
+
+The code generated for 8 and 16-bit integer types is generally not as
+efficient as the code generated for 32-bit integer types.  It is generally
+worthwhile to use 32-bit integer types for intermediate computations, even
+if the final result will be stored in a smaller integer type.
+
 Low-level Vector Tricks
 -----------------------
 

expr.cpp

@@ -1659,7 +1659,7 @@ AssignExpr::TypeCheck() {
     if (rvalue != NULL) 
         rvalue = rvalue->TypeCheck();
     if (rvalue != NULL && lvalue != NULL) 
-        rvalue = rvalue->TypeConv(lvalue->GetType(), "operator =");
+        rvalue = rvalue->TypeConv(lvalue->GetType(), "assignment");
     if (rvalue == NULL || lvalue == NULL) 
         return NULL;
 

stdlib.ispc

@@ -2363,7 +2363,7 @@ static inline uniform float half_to_float(uniform unsigned int16 h) {
         uniform unsigned int32 hs = h & (int32)0x8000u;  // Pick off sign bit
         uniform unsigned int32 he = h & (int32)0x7C00u;  // Pick off exponent bits
         uniform unsigned int32 hm = h & (int32)0x03FFu;  // Pick off mantissa bits
-        cif (he == 0) {  
+        if (he == 0) {  
             // Denormal will convert to normalized
             uniform int e = -1;
             // The following loop figures out how much extra to adjust the exponent
@@ -2485,7 +2485,7 @@ static inline uniform int16 float_to_half(uniform float f) {
             ret = (xs >> 16);
         } 
         else {
-            cif (xe == 0x7F800000u) {  
+            if (xe == 0x7F800000u) {  
                 // Inf or NaN (all the exponent bits are set)
                 if (xm == 0)
                     // Zero mantissa -> signed infinity