Hello,

im trying to get a compute shader to to execute with localsize.x = 1024, localsize.y = 10, localsize.z=1. However I get the following error when trying to compile: Total number of invocations specified by GROUP_SIZE exceeds implementation limit.

I queried the limitations of my hardware and they are: x=1536, y=1024, z=64; x * y * z=37099. So if I’m not mistaken my specified groupsize should be possible?

Through tests I found that everything compiles fine as long as I stay within x=1024, y=1024, z=64; x * y * z=1024

I appreciate any help!