Justin Szaday

Hi, I'm on the `tf.distribute` team. I'll start looking into this issue, and reply once I have insights into it. *UPDATE 1:* It seems like the user-specified strategy and its...

After looking into this extensively, I realized that during the model-saving process, a `tf.function` is run with the default replica context and strategy, and the output from its `tf.print` statements...

Yeah, the real problem is the `tf.print` as far as I can tell. And, as for your logs, those look favorable to me; others can weigh in if there are...

I am unsure about how to compile MSA with the `build` script. I typically run `make` from `src/libs/ck-libs/multiphaseSharedArrays/` to make `-module msa` available.

Hello Geoffrey, MPI-based Charm++ builds do not seem to exhibit these problems on an M1. After a quick: `brew install openmpi` Then: `./build charm++ mpi-darwin-arm8 --with-production -g3 -j` I was...

Good point. I checked v7.0.0 and it performs comparably on my Mac. I can, at least, demonstrate poor scaling for `netlrts` and `multicore` builds, so I'll focus my efforts there....

Thank you for bringing this to our attention. These instructions have become out of date with the switchover to the `cmake`-based build system. I will update them.

It's odd that it's using `clang` on `test.cpp`; not certain about why. I know that Charm++ determines which compiler to build `charmxi` based on the platform, and that bypasses the...

It believe that PR #3475 changed the default assumptions of `charmc`. Future versions (after v7.0.0) will now require passing flags like `-c` and `-S` explicitly. Also, you may not need...

Alright, duly noted. I was not aware those are a thing and my editor (Atom) just auto inserts HTML-style comments when using the comment keyboard macro (Ctrl+/); perhaps I should...