2D threads & blocks In what order they're executed?

Hi All!

For example I run the following kernel:

KERNEL<<<dim3(BLOCKS_X, BLOCKS_Y), dim3(THREADS_X, THREADS_Y)>>>(some params);

Could you please describe an order of that threads & blocks execution?

The execution order should be considered random. It’s completely up to what the scheduler thinks would be best for load balancing, you should not assume any specific order.