Mark Taylor

@bartgol can you double check this? This bug is referring not to file offsets, but array indexes in init_decomp. Back in 2022, PIO did not support using int64 for these...

Per the original issue, #2698, its the GLL restart array which will break at NE6144. Note that even with PG2 physics, dynamics variables are on np4 (unique points representation) and...

This issue was created in 2022 because the atmosphere restart arrays, at NE=6144 will exceed 2,147,483,647 These arrays have an index of size: 6144\*6144\*6\*16 = 3623878656 I found an even...

For MPAS, is config_dt = 900.0 the coupling interval or the timestep? Timesteps should appear in the namelist. But coupling intervals should not, to avoid confusion with the coupling interval...

@rljacob : so on the confluence page that pointed to this issue, it would be best to leave config_dt in the MPAS namelist, but remove 'dtime' from both atmosphere and...

it's dangerous, so I just want to be sure the practice does not get archived in confluence!

I think there may be some confusion here. @jsbamboo is wanting to output remapped dynamics grid data, and has a np4/gll map file to the dst grid. @bartgol seems to...

There must be code already to convert data from the np4 DG representation to the uniquepoints representation. It most likely is used when running np4 physics, and also when outputing...

Sorry, I completely missed that. Way too much abstraction of the "remap" operation for me :-)

if this check fails, doesn't it means there is some real data corruption somewhere, usually in the MPI systems. I would worry that means its not safe to continue the...