Joshua Patterson

Despite the fact that this is simply not true, numerous companies have adopted RAPIDS and many startups depend on RAPIDS. Most companies have SW audit process to make sure they...

What's not true is "no company" I think that's a stretch. On the other side, RAPIDS is not just a single library, it's a bridge of many libraries. For instance,...

I'm arguing that our adoption of conda alleviates many other issues, we wouldn't want to move from conda, and companies are adopting RAPIDS. Having a way to isolate GPL libraries...

> If they knew no company could easily adopt rapids because of conda mixing in GPL, they might not have made that jump. My *only* point is you can't say...

We will just have to agree to disagree. The facts of your argument are strong enough to stand on their own. Stating an opinion that NVIDIA naively made a decision...

I strongly agree conda being easier would be beneficial. This is why NVIDIA funds conda-forge and has people contributing to it. It would be great if others did the same.

Given this is built on Dask, RAPIDS integration "could" be somewhat straight forward to see if acceleration is of value.

Given this seems ugly in general, and I hate to say let's create yet another standard, but should we start with OpenMPI and then slowly add other variants as time...

> To be clear, the APIs I described are part of _the_ MPI standard that all MPI implementations _should_ implement, it's not specific to OpenMPI. > > From what I...

> As a trial: I've made a short google form that looks like > >  > > That'll be dumped into a...