Hi Amy,
Congrats on the new computer and great to hear you are doing an analysis update.
I'm heads down right now but should have some time next week to look at lne 763 in the code and get better recommendations. Or you can in an standard python reader or on the github page.
In the meantime, my initial thoughts:
1) 700 links and 1.3 billion cells will be tough, but you have a souped up machine so that should be fine, and this is the same or similar numbers as last time I think... so that doe snot seem to be it. especially since it is crashing on the first try at this step...
2) Often times new Esri releases break some aspect of the code. It could be something in the new ArcPro version. It came out in late May it looks like, and it is possible that you are the first person coming across a configuration that triggers the new bug. But.. you said it worked with a small test region, so maybe it is only particular size or overlay characteristics that trigger the bug. Unlikely but possible. You could roll back to a previous version, or even last year's version to be on the safe side there...
3) given the above, and that it is having a hard time building the attribute table, on some of the error messages but not all, with all cells, c:\lma2\scratch\core1600\addras have NODATA value , maybe there is something with the way your resistance surface and cores (hubs right?) are configured...When you reply, please clarify if you tried it for multiple subregions and it gets stuck in the same spot. And if different errors, please include those log files too.
Thanks,
JOhn