Actually it seems the problem is something else. The following code locks up my system even if I run it over a global and local work size of (1,1). Can someone tell me what the heck is wrong with this kernel. I had a kernel that was working and I modified it and I can't find out what the heck is wrong so I boiled it down until just the following. I am not using one of the input images but it doesn't seem to matter. Maybe I am just having one of those days NOTE: I may have made a typo because the code is on a machine that isn't on the network so I had to retype it but the following is the basic code:
__kernel void Example(__read_only image2d_t imgA, __read_only image2d_t imgB, __write_only image2d_t imgC, sampler_t sampler, int width, int height)
{
int column = get_global_id(0);
int row = get_global_id(1);
int2 coords = (int2)(column, row);
int value;
value = read_imagei( imgA, sampler, coords ).x;
if(row < height && column < width) {
write_imagei( imgC, coords, value );
}
}
