Closed aseyboldt closed 1 month ago
that sounds strange since the algorithm is supposed to be deterministic (though i might have messed up somewhere). let me know if you manage to reproduce it
i wrote a couple fixes for the svd/eigendecomp to make the root finder more robust. i'll publish a patch release after i finish a couple more stuff can you check if the code is still panicking if you use the main branch?
Seems like this fixed the issue, thank you. (And sorry for the bad issue, I'll do better if there is a next time ;-) )
Am Do., 22. Aug. 2024 um 02:25 Uhr schrieb sarah @.***>:
i wrote a couple fixes for the svd/eigendecomp to make the root finder more robust. i'll publish a patch release after i finish a couple more stuff can you check if the code is still panicking if you use the main branch?
— Reply to this email directly, view it on GitHub https://github.com/sarah-ek/faer-rs/issues/141#issuecomment-2303396313, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAOLSHLXF4ZTXLPOEDU7NC3ZSU4YDAVCNFSM6AAAAABKQJ2VLGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMBTGM4TMMZRGM . You are receiving this because you authored the thread.Message ID: @.***>
fixed in 0.19.2
Describe the bug I get very sporadic panics when I compute selfadjoint eigenvalue decompositions from here: https://github.com/sarah-ek/faer-rs/blob/main/src/linalg/evd/tridiag_real_evd.rs#L349
To Reproduce I have trouble reproducing this, I think it might not always happen with the same matrix, and only very rarely. Hopefully I can add a matrix that showed this behavior when it happens again.
Expected behavior no panic
Desktop (please complete the following information): arch linux