CITATION.cff

cff-version: 1.2.0
message: "If you use IntelliPerf or discuss our work in your research, please always cite our work."
authors:
  - family-names: "Awad"
    given-names: "Muhammad"
  - family-names: "Ramos"
    given-names: "Cole"
  - family-names: "Lowery"
    given-names: "Keith"
title: "IntelliPerf: LLM-Powered Autonomous GPU Performance Engineer"
doi: "10.5281/zenodo.15845118"
date-released: 2025-07-08
url: "https://github.com/AMDResearch/intelliperf"
repository-code: "https://github.com/AMDResearch/intelliperf"
license: MIT
keywords:
  - "GPU optimization"
  - "performance engineering"
  - "machine learning"
  - "LLM"
  - "ROCm"
  - "AMD"
  - "automated optimization"
  - "bank conflicts"
  - "memory access patterns"
  - "atomic contention"
abstract: "IntelliPerf is an automated performance engineering framework that addresses the complex challenge of GPU kernel optimization. It systematizes the optimization workflow by orchestrating a comprehensive toolchain that automatically profiles applications using rocprofiler-compute, identifies high-level bottlenecks with Guided Tuning, pinpoints specific source code lines using Omniprobe, generates optimized code through Large Language Models (LLMs), and validates results using Accordo for correctness and performance."