Closed Flash-LHR closed 1 year ago
Hi,
Thanks! Is this related to this? https://github.com/meelgroup/bosphorus/issues In other words, does that PR fix this bug or it causes this bug? I understand that this needs fixing, but want to understand first how to reproduce.
Cheers,
Mate
This issue has nothing to do with pr.
Pr only sets the number of threads of cryptominisat5. Before pr, the number of threads of cryptominisat5 is always 1 regardless of whether the user specifies the thread parameter. In fact, I did not attempt to fix the above assertion failure.
I have uploaded the file r05.cnf for reproducing.
Hi,
Sorry for the late response. It was indeed a bug! I fixed it in CryptoMiniSat, here: https://github.com/msoos/cryptominisat/commit/9b4ee6f14849f7ea2b9efe2f33b698a3ba203947
If you recompile both CMS and Bosphorus (maybe no need to recompile Bosphorus, depending on your setup), it should now work! I have tested it :)
I would like to thank you for both the perfectly reproducible problem description and your patience. Sorry, lately I've been very busy so I couldn't get to this as soon as I hoped. Nevertheless, about 1 week to fix is probably not too bad. Thanks again for the bug report, it really helped,
Mate
Thank you very much, it helps me a lot.
Besides, I have made a pr to set threadNum for cryptominisat in bosphorus, maybe you can consider that.
Ah, good point, thanks for the reminder! I merged the PR as well, and cleaned up the code a bit along the way.
The operation succeeds when using the default number of threads, but the assertion fails after setting the number of threads. OS: 5.17.0-kali3-amd64 CPU:Intel(R) Core(TM) i5-8300H CPU @ 2.30GHz
r05.cnf.txt