Bradley Dice

icon indicating a website link https://bradleydice.com icon indicating an email [email protected]

icon company @NVIDIA @rapidsai icon location Kansas City, MO icon location Scientific software developer, working on @rapidsai cuDF and RMM. Python, C++, CUDA, algorithms & data. PhD in Physics & Scientific Computing from @glotzerlab.

Results 608 comments of Bradley Dice

[FEA] Improve Null-Aware Operator Support in AST-Codegen

I've asked for @lamarrr to split the segfault fix into a separate PR before we merge.

Add RMM User Guide

I've broken this PR into a few parts that are ready to merge: - #2137 - #2138 - #2139 After #2137 merges, I will start breaking up the new user...

[FEA] Properly support alignment argument in `do_allocate` / `do_deallocate`

I discussed with @mhaseeb123. The premise of this issue might be flawed. We are moving away from having a base class with the new CCCL MR design, so it might...

Use pytest-xdist to parallelize tests

/ok to test

CCCL 3.2 resource ref compatibility

/ok to test

Make UDF source type parameter strongly-typed

@lamarrr See libcudf developer docs on deprecation: https://github.com/rapidsai/cudf/blob/branch-25.08/cpp/doxygen/developer_guide/DEVELOPER_GUIDE.md#deprecating-and-removing-code

Test with latest CCCL

/ok to test

Test with latest CCCL

/ok to test

Test with latest CCCL

@miscco It seems like some issues are resolved but there are still build problems. I'm not sure if this is RMM or CCCL changes at fault. I will try to...

[FEA] Support memory resources from CCCL 3.2

Current plan for adoption: ### Allocation Interfaces This list of tasks requires CCCL 3.1+, so we can ship these changes in 25.12. 1. Build RMM w/ 3.1 via polyfill and...