William R Tobin

Is this essentially a duplicate of #707 ?

I'll modify #1098 to include removing the pointer-based packing and instead use the array1d type.

No. I think we wanted to also work on the overall call-stack for the packing functions starting from the group/wrapper-level to just refactor the entire thing at once, but haven't...

Note I haven't checked for all tests passing on Lassen yet -- since the machine is still down at the moment.

Caliper was updated in the TPLs prior to the initial time-hist functionality merge, but I don't think the caliper function-marking macro to account for overloaded functions and multiple instantiations of...

While working on this it would probably be good to re-instrument the packing functions now that they're largely async, in order to more accurately reflect their performance, as the current...

> @wrtobin do we have an associated TPL PR? @TotoGaz Not yet as we're starting to move to Spack and spack environments to do our TPL build and install and...

Yeah we'd decided to not provide official support for non-MPI builds. At the same time we decided not to allocate resources to obsolete our previous support for the non-MPI build....

There was some motivation coming from an error on frontier, but IIRC we've since moved past that. Regardless, there is never a reason (and should never be, by design) we...

Randy also pointed to streak (hardware-wise its essentially 1 node of lassen we're working using as a CI runner) which is having cuda memory errors with lvarray string arrays: https://github.com/GEOS-DEV/GEOS/actions/runs/7674651265/job/20919616872?pr=2873...