Is there any "pragma unroll" to control the loops as CUDA does?
thx
Currently no.
Can I use the mod op for this?
Example:
void myFunc ( const int n )
{
int i;
for ( i=0; i<n%16; ++i ) // %16 because "n" should never be greater than 16
....
}
You should not be facing any issue with this.
