Roger Pawlowski
Roger Pawlowski
We are very careful about this in the nox solvers and make sure to recalculate norms (instead of reusing across Newton iterations) since weights can change even between Newton steps....
@e10harvey - assigning to @dridzal
Since there are a bunch of packages disabled in the cuda 11 build (STK, NOX, Panzer at least), we can't drop the cuda 10 build yet until they are covered....
@bartlettroscoe - David is working on updating the empire builds to c++17 and new TPLs stacks as we speak. The plan is to have it done near the end of...
@jhux2 The current use of vortex was a stopgap measure to recover PR testing. When weaver burned out, vortex was the only option at the time. It was intended to...
@bartlettroscoe @csiefer2 - I've got the cuda 11 errors figured out. The cuda 11 build is linking against a blas/lapack that is threaded and built for openmp. The new build...
> Do we have any documentation for that? Before these changes are pushed into Trilinos, we will have instructions and an example in Trilinos (probably in the demos directory). It's...
@cgcgcg - this is not quite ready for review yet, but feel free to review it now if you want.
> What if you just had an option as to whether to adjust the step-size to hit t_final? I thought the option to hit a final time point exactly already...