We won't benchmark with this case, since I'm skeptical it will outperform cacqr with complete_inv=true, but I'd like to have this working nonetheless. I tried to get it working, but it seemed to fail, although I checked the Q factor element by element on rank 0 and it seemed to be correct (it matched the correct one). Perhaps there was an error on a different process?
We won't benchmark with this case, since I'm skeptical it will outperform
cacqrwithcomplete_inv=true, but I'd like to have this working nonetheless. I tried to get it working, but it seemed to fail, although I checked the Q factor element by element on rank 0 and it seemed to be correct (it matched the correct one). Perhaps there was an error on a different process?Here is a picture of the code I was working with: