aaron/ispc
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

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

Browse Source 
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.
This commit is contained in:
Matt Pharr
2013-07-23 17:01:03 -07:00
parent 9ba49eabb2
commit f7f281a256
61 changed files with 166 additions and 120 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/exclusive-scan-and-2.ispc
View File

@@ -3,7 +3,7 @@ export uniform int width() { return programCount; }
export void f_f(uniform float RET[], uniform float aFOO[]) {
RET[programIndex] = -1;
int32 a = ~(1 << programIndex);
int32 a = ~(1ul << programIndex);
if ((programIndex < 32) && (programIndex & 1) == 0) {
RET[programIndex] = exclusive_scan_and(a);
}
@@ -15,7 +15,7 @@ export void result(uniform float RET[]) {
if ((programIndex & 1) == 0 && programIndex > 0 && programIndex < 32) {
int val = 0xffffffff;
for (int i = 0; i < programIndex-1; i += 2)
val &= ~(1<<i);
val &= ~(1ul<<i);
RET[programIndex] = val;
}
}