New implementation of CU mask testing that focuses on correctness of masking. Unlike previous implementations, this new implementation does not rely on performance measurements to decide on the results of the test. Instead, this implementation checks if waves were executed on all the CUs enabled and only the CUs enabled. Test case initially supported on GFX12. Signed-off-by: David Belanger <david.belanger@amd.com> Change-Id: I5af8b890179bc9a415fc7f47e736f4971fc40c4a