Reproducible memout + nondeterminism #47
Description
Dear Authors of Quokka-Sharp-ECMC,
Our group is working on a benchmarking study for quantum circuit equivalence checking. In our pipeline, we generate Clifford+T circuits, optimize them with PyZX, and then use quokka_sharp_ECMC to check equivalence.
We have encountered a set of circuit patterns that occasionally produce non-deterministic outcomes on my machine (e.g., q32_depth33_index_64 in the attached bundle). In earlier runs, the same circuit sometimes returned different results (True/False). In recent reruns, the behavior has shifted: the runs now fail consistently with MEMOUT, and the per-circuit results no longer fluctuate.
I’m attaching a minimal reproduction package (README + relevant inputs/logs). The bundle includes several circuit patterns that trigger errors on one of the two machines I tested; some patterns reproduce the issue on both machines. Given the MEMOUT failures and the earlier run-to-run variability, I suspect this may be related to internal memory management rather than misuse on our side.
Could you please take a look and let me know whether I might be missing something, or whether this points to an internal issue in quokka_sharp? Thanks very much!
Best Regards,
runtime_q32_d33_raw_that_partially_works.txt
Xuan Du Trinh