llnl / llnl/CompilerGPT

potential for regressions in the iterative process

Open
#16 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
C
Stars
19
Forks
1
PR merge metrics
No merged PRs in 30d

Description

There’s a risk that the LLM could generate a change that worsens performance or compiler optimizability in ways that aren’t immediately obvious. The iterative nature means if performance doesn’t improve, the loop continues, but the LLM isn’t directly told the previous iteration’s performance was worse – only the absolute execution time. It might infer it if the time went up, but the prompts mainly emphasize further reduction.

The framework doesn’t explicitly rollback to a previous better version; it just keeps modifying. In theory, it could oscillate or even end up with no net improvement, which might waste time.

Contributor guide

No contributing guide indexed for this repository

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

No files, tests, or entry points are named. Start by locating the iterative optimization loop in CompilerGPT and inspect how execution time is passed between iterations. Done should be defined around retaining the best-known result and handling regressions or oscillation, with tests covering the iteration behavior.

Written by the indexing model from the issue text.

Assessment

Tech stack
c
Domain
ai, compilers, performance
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.