You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit b270d2c
Browse filesBrowse the repository at this point in the historyBrowse files
feat: fall back to the default solver for each coordinate, not each feature
The coordinates 1 and 2 of a fault frame have orthogonal rows that the
native core cannot read, so all the faults used cg. Now the plugin finds
the coordinates that the core refuses (with the support check of
loop-interpolation), solves only those with the default solver, and keeps
native for the others. It records the coordinates, so the next build does
not try native for them again. If the check finds no coordinate, the full
feature uses the default solver, as before.
The tests for the ADMM solver of inequality rows now turn the native core
off, because with it the native solver solves the inequalities.
0 commit comments