NVIDIA / NVIDIA/cudf

[FEA] Explore lto fragments in cudf-polars

Open
#24,200 0 comments 0 reactions 0 assignees View on GitHub
cudf-polars feature request
Dominant language
C++
Stars
9.8k
Forks
1.1k
Avg merge
3d 6m
Merged PRs (30d)
278

Description

libcudf now offers facilities to pre-bake "LTO fragments" via jit or similar and then load and use them at runtime.

In cudf-polars we have an increasing number of custom AST transformations that are hand-coded. We might imagine that turning these into LTO fragments (given we know what types might be needed ahead-of-time) would be performance-beneficial

Contributor guide

Open the contributing guide

Research direction

Start by reading libcudf's LTO fragment facilities and the hand-coded custom AST transformations in cudf-polars. Determine whether known type requirements allow those transformations to be represented as runtime-loaded LTO fragments and whether that is performance-beneficial; document the feasibility and measured outcome.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
data, performance
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Active
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.