NVIDIA / NVIDIA/cudf

[PERF]: cuCascade byte range prefetch backend for hybrid scan

Open
#23,536 0 comments 0 reactions 0 assignees View on GitHub
cudf-polars Performance
Dominant language
C++
Stars
9.8k
Forks
1.1k
Avg merge
3d 6m
Merged PRs (30d)
278

Description

Experimental byte range prefetch backend using cuCascade's reactor-based IO engine. Uses fadvise hints to trigger downloads into cuCascade's internal prefetch cache.

PR: #23450

Contributor guide

Open the contributing guide

Research direction

Start with PR #23450, since the issue identifies it as the implementation location; no source files, tests, or entry points are named here. Review how the proposed reactor-based IO engine uses fadvise hints and cuCascade's internal prefetch cache, then confirm that the hybrid scan backend provides the described byte-range prefetch behavior and its validation.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
backend, performance
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.