Vyas Ramasubramani
Vyas Ramasubramani
Is there a long-term plan to pull more of the concrete implementations from rmm into CCCL? That seems like the best way to broaden adoption and usage of these allocators...
I would like to add equality checks, but I'm not comfortable with any tolerance-based checks. I think only _exact_ equality should match, just like with numpy arrays or similar. We...
One alternative would be to simply remove the base class implementations. Having these "abstract" (in quotes because they aren't really abstract in the sense of Python's `abc` module) methods and...
This will be review for most people discussing here, but just wanted to summarize explicitly to make sure we're clear and all on the same page: ### Status quo -...
I don't think that's a major concern. In theory I suppose it could happen from copy-pasting top-level CMakeLists.txt files, which is I imagine how many projects that aren't RAPIDS projects...
You're right that libcudf uses it, although it does so a bit differently since both cuml and raft both support configuring whether it's called using the `DETECT_CONDA_ENV` option and they...
Seems closely related to https://github.com/rapidsai/cudf/issues/15825, just for a different format, right? Just wondering if it's worth rolling that change in too; no worries if not.
/ok to test
> > I was wondering if we need to update corresponding python bindings in `python.pxd` and perhaps `python.pyx` as well? > > @mhaseeb123 I was worried about that when I...