rhdodds
commented
9 months ago This is bug came about when I changed the default, minimum number of global Newton iterations to 2 from 1.
The code is now fixed to work properly for the case when the user constrains all model displacements and does not set the minimum number of iterations = 1.
Until I release new executables on www.warp3d.net http://www.warp3d.net/, the work around is to set the minimum number of iterations to 1.
E.g.
nonlinear analysis parameters solution technique direct sprase maximum iterations 5 minimum iterations 1 convergence test norm res tol .01 ….. ….
On Feb 2, 2024, at 12:44 PM, Mark Messner @.***> wrote:
This seems to work to fix the bug I found on Linux, but obviously needs more testing before merging in.
You can view, comment on, or merge this pull request online at:
https://github.com/rhdodds/warp3d/pull/22
Commit Summary
e46793a https://github.com/rhdodds/warp3d/pull/22/commits/e46793ac6b402e621cb502da6743775e35973bdc Simple fix for avoiding pardiso segfaults if we don't actually have equations File Changes (1 file https://github.com/rhdodds/warp3d/pull/22/files) M src/drive_assemble_solve.f https://github.com/rhdodds/warp3d/pull/22/files#diff-b094572b62c8809117167845d630fd5858a9ac1041c8f82a06a424d21130bcd0 (5) Patch Links:
https://github.com/rhdodds/warp3d/pull/22.patch https://github.com/rhdodds/warp3d/pull/22.diff — Reply to this email directly, view it on GitHub https://github.com/rhdodds/warp3d/pull/22, or unsubscribe https://github.com/notifications/unsubscribe-auth/AC7CB7UPJE3Y6RT2SPI5A53YRU6ZJAVCNFSM6AAAAABCXGPDCKVHI2DSMVQWIX3LMV43ASLTON2WKOZSGEYTKNRVGIYTANA. You are receiving this because you are subscribed to this thread.
This seems to work to fix the bug I found on Linux, but obviously needs more testing before merging in.
Would fix #21