Piyush Sharda comments

Results 72 comments of


                                            Piyush Sharda

consider SciPy BDF solver as alternative to VODE

+1 for this idea. BDF works well for the primordial_chem stuff too.

refuse to build when Cray compiler is detected for CPU-only builds

So this is an issue if you compile using `setonix-cpu.profile` and `setonix-gpu.profile`?

create a struct to hold the Microphysics runtime parameters

Looks like we will also need to modify CMakeLists to reflect these changes. @BenWibking ?

amrex::Abort does not unconditionally abort on GPU

Fixed for PopIII by #575

Mask Microphysics burn for covered cells

As an example, chemistry is taking 98% of the time in a PopIII simulation run on (AMD) Setonix GPUs with 14 AMR levels, and 36 million total cells: ``` ---------------------------------------------------------------------------------------------------...

in rare cases, NSCBC outflow produces unphysically high velocities

@BenWibking is this now resolved by #598 ?

enable FPE traps in CI

/azp run

docs: add bullet point list of new features

Write a code paper and call it Quokka 2.0? ;)

Piyush Sharda