Recently on both testing and main site branches of CPDN It has come to the attention of both myself and the team at Oxford that tasks that are virtually all running without issue under a WINE installation of BOINC on Ubuntu 23.04 are nearly all failing on real Windows boxes. One of the scientists involved believes these tasks will not in fact be useful to the project as the segfault is in all probability due to the physics of the model going out of control.

Does anyone who knows more about WINE than I do (My knowledge is limited to being able to install it and run Windows programs using it.) know if anything has change with recent iterations of WINE that might cause this behaviour?