NVIDIA / NVIDIA/cuCollections

[ENHANCEMENT]: Switch to cuda::memory_resource once it is available

Open
#289 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

P3: Backlog type: feature request
Dominant language
Cuda
Stars
667
Forks
120
Avg merge
6d 7h
Merged PRs (30d)
4

Description

Is your feature request related to a problem? Please describe.

We currently roll our own default cuco::cuda_allocator, which internally calls cudaMalloc/cudaFree.
This approach doesn't leverage the concept of stream-ordered allocations, which might degrade performance for operations such as size() and insert(), where we allocate intermediate storage to retrieve the count.

Describe the solution you'd like

libcu++ v2.0 introduces a new cuda::memory_resource (design, initial PR, final PR).

We should use this facility instead.

Describe alternatives you've considered

No response

Additional context

No response

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by locating cuco::cuda_allocator and its uses in the implementations of size() and insert(). Check whether the referenced libcu++ v2.0 cuda::memory_resource facility is available, then assess the allocator switch while preserving intermediate-storage behavior and performance. Done means the custom allocator is replaced where appropriate and the affected operations remain correct.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
performance
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
32/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.