Graham Markall
Graham Markall
To answer the question generally - we update to newer versions of LLVM as fast as we are able, but given resource constraints that's not particularly fast. In addition to...
> Generally speaking, what is the benefit to numba users for upgrading the bundled LLVM? Do these usually come with performance improvements for compiling user code & run time of...
Requires a merge from `main` to resolve the RTD issue.
(I am working with @Matt711) @jeanas Many thanks for reviewing this - this PR was opened prior to us receiving approval to contribute it upstream to Pygments due to a...
We received permission / approval to make this contribution upstream, so we can now continue to work on this PR - many thanks for your time and understanding so far...
What happens if a user writes something like ```python @cuda.jit def make_full_name(first_name, surname): return first_name.concat(" ").concat(surname) def names_udf(df): return make_full_name(df['first_name'], df['surname']) full_name_series = df.apply(names_udf) ``` Apologies for the rustiness of...
One thing that's becoming apparent to me in reviewing this is that it's very hard to understand the idea of the design just by reading the code, even though I...
My understanding is that you're run tests with `compute-sanitizer` and things seem to appear OK from that perspective. I think we have NRT stats available in Numba-CUDA now - do...
xref also to https://github.com/rapidsai/cudf/issues/13351
I'm not sure this needs triage or going stale, so I'm going to remove both.