severity 1103553 wishlist tags 1103553 upstream On Sun, Apr 20, 2025 at 12:21 AM Santiago Vila <sanv...@debian.org> wrote: > > El 19/4/25 a las 20:32, Roman Lebedev escribió: > >> The cpuinfo is attached. > > Right, so that CPU does claim to have AMX support. > > But is AMX supported in such a virtualized environment? > > https://community.intel.com/t5/Intel-Tiber-Developer-Cloud/Can-t-run-AMX-Instructions-on-Intel-Xeon-4th-Gen-Scalable/m-p/1532428 > > suggests not. Could you please file an issue with AWS? > > I really feel like the issue is on their end... > > Maybe it is, but it is not easy for me to file an issue with them, > because I'm just using an account for which I have limited permissions > without being the customer myself. > > Given that it's a single test the one that fails, I was thinking about > disabling it with a changelog entry like this: > > "Temporarily disable test foo, since it does not seem to work in > some hardware configurations." This feels like disabling tests because they happen to fail because of bad ram. Because that is exactly what is going on - the system is knowingly lying about the CPU features it provides.
> Would you accept such a patch? I am not enthusiastic about breaking the package's test suite any further. > (I honestly think it would be a lot better than nothing). > > Thanks.
