Choose type for integer literals to match the target mask size (if possible).

On a target with a 16-bit mask (for example), we would choose the type
of an integer literal "1024" to be an int16.  Previously, we used an int32,
which is a worse fit and leads to less efficient code than an int16
on a 16-bit mask target.  (However, we'd still give an integer literal
1000000 the type int32, even in a 16-bit target.)

Updated the tests to still pass with 8 and 16-bit targets, given this
change.

tests/aossoa-2.ispc

@@ -2,8 +2,8 @@
 export uniform int width() { return programCount; }
 
 export void f_v(uniform float RET[]) {
-#define width 4
-#define maxProgramCount 64
+#define width 4ul
+#define maxProgramCount 64ul
     assert(programCount <= maxProgramCount);
 
 //CO    const uniform int width = 4;