Basit Ayantunde

I need this too

I'd like to take on this once triaged

@StephanTLavavej I'd like to take on this if it's still available

> Is there a particular way you would suggest reporting cache hit rates via NVTX? As long as the NVTX regions are added for the entire compilation and caching process,...

> I'm guessing they will want to leave this kind of control up to the calling application as well. I agree with @davidwendt here, the user would have a better...

re-pinging @bdice

Presently blocked by a change to the preprocessing method in JITIFY. The new method loads all included headers even if they are guarded by macros that disable them. It is...

succeeded by https://github.com/rapidsai/cudf/pull/19561

A per-thread `arena_memory_resource` or `pool_memory_resource` (optionally with a fallback to a shared upstream) would be ideal when performing multi-threaded allocations. There will always be contention and synchronization overhead as long...

/ok to test 7f97737