Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ArchAgent: Agentic AI-driven Computer Architecture Discovery

Raghav Gupta, Akanksha Jain, Abraham Gonzalez, Alexander Novikov, Po-Sen Huang, Matej Balog, Marvin Eisenberger, Sergey Shirobokov, Ngân Vũ, Martin Dixon, Borivoje Nikolić, Parthasarathy Ranganathan, Sagar Karandikar

TL;DR

This work shows ArchAgent's ability to automatically design/implement state-of-the-art cache replacement policies (architecting new mechanisms/logic, not only changing parameters), broadly within the confines of an established cache replacement policy design competition, and outlines broader implications for computer architecture research in the era of agentic AI.

Abstract

Agile hardware design flows are a critically needed force multiplier to meet the exploding demand for compute. Recently, agentic generative AI systems have demonstrated significant advances in algorithm design, improving code efficiency, and enabling discovery across scientific domains. Bridging these worlds, we present ArchAgent, an automated computer architecture discovery system built on AlphaEvolve. We show ArchAgent's ability to automatically design/implement state-of-the-art (SoTA) cache replacement policies (architecting new mechanisms/logic, not only changing parameters), broadly within the confines of an established cache replacement policy design competition. In two days without human intervention, ArchAgent generated a policy achieving a 5.3% IPC speedup improvement over the prior SoTA on public multi-core Google Workload Traces. On the heavily-explored single-core SPEC06 workloads, it generated a policy in just 18 days showing a 0.9% IPC speedup improvement over the existing SoTA (a similar "winning margin" as reported by the existing SoTA). ArchAgent achieved these gains 3-5x faster than prior human-developed SoTA policies. Agentic flows also enable "post-silicon hyperspecialization" where agents tune runtime-configurable parameters exposed in hardware policies to further align the policies with a specific workload (mix). Exploiting this, we demonstrate a 2.4% IPC speedup improvement over prior SoTA on SPEC06 workloads. Finally, we outline broader implications for computer architecture research in the era of agentic AI. For example, we demonstrate the phenomenon of "simulator escapes", where the agentic AI flow discovered and exploited a loophole in a popular microarchitectural simulator - a consequence of the fact that these research tools were designed for a (now past) world where they were exclusively operated by humans acting in good-faith.

ArchAgent: Agentic AI-driven Computer Architecture Discovery

TL;DR

This work shows ArchAgent's ability to automatically design/implement state-of-the-art cache replacement policies (architecting new mechanisms/logic, not only changing parameters), broadly within the confines of an established cache replacement policy design competition, and outlines broader implications for computer architecture research in the era of agentic AI.

Abstract

Agile hardware design flows are a critically needed force multiplier to meet the exploding demand for compute. Recently, agentic generative AI systems have demonstrated significant advances in algorithm design, improving code efficiency, and enabling discovery across scientific domains. Bridging these worlds, we present ArchAgent, an automated computer architecture discovery system built on AlphaEvolve. We show ArchAgent's ability to automatically design/implement state-of-the-art (SoTA) cache replacement policies (architecting new mechanisms/logic, not only changing parameters), broadly within the confines of an established cache replacement policy design competition. In two days without human intervention, ArchAgent generated a policy achieving a 5.3% IPC speedup improvement over the prior SoTA on public multi-core Google Workload Traces. On the heavily-explored single-core SPEC06 workloads, it generated a policy in just 18 days showing a 0.9% IPC speedup improvement over the existing SoTA (a similar "winning margin" as reported by the existing SoTA). ArchAgent achieved these gains 3-5x faster than prior human-developed SoTA policies. Agentic flows also enable "post-silicon hyperspecialization" where agents tune runtime-configurable parameters exposed in hardware policies to further align the policies with a specific workload (mix). Exploiting this, we demonstrate a 2.4% IPC speedup improvement over prior SoTA on SPEC06 workloads. Finally, we outline broader implications for computer architecture research in the era of agentic AI. For example, we demonstrate the phenomenon of "simulator escapes", where the agentic AI flow discovered and exploited a loophole in a popular microarchitectural simulator - a consequence of the fact that these research tools were designed for a (now past) world where they were exclusively operated by humans acting in good-faith.
Paper Structure (47 sections, 9 figures, 2 tables)

This paper contains 47 sections, 9 figures, 2 tables.

Table of Contents

  1. Introduction
  2. The ArchAgent system
  3. Automatically designing state-of-the-art cache replacement policies
  4. Policy deep-dive
  5. Designing beyond SPEC
  6. Post-silicon "hyperspecialization"
  7. Strengths, pitfalls, and a call-to-action
  8. Background
  9. LLM-based Evolutionary Agents
  10. Cache Replacement Championships
  11. ArchAgent System Design
  12. Prompting an LLM to be an Architecture Research Agent
  13. Starter Code
  14. Workloads and ChampSim Hardware Configurations
  15. Evaluation Metrics
  16. ...and 32 more sections

Figures (9)

  • Figure 1: High-level system diagram of ArchAgent, our agentic-AI-based computer architecture discovery system. In this example, novel cache replacement policy candidates are automatically designed/implemented by AlphaEvolve in ChampSim, a popular trace-based microarchitectural simulator. ChampSim is then compiled and run with a specified workload suite (e.g., SPEC) to evaluate the new policy on a target metric (e.g., IPC). This process continues iteratively, with ArchAgent continually proposing and evaluating new logic/mechanisms within the policy.
  • Figure 2: Simplified example of a prompt given to the AlphaEvolve used in ArchAgent including persona, background information, guidance.
  • Figure 3: Example Policy31 modifications in the form of a diff to implement the Hawks and Doves mechanism. The packed_usage_counter and corresponding get-, increment-, and reset_usage setter/getters are used to help determine eviction candidates.
  • Figure 4: Performance improvement (suite-level geomean IPC speedup normalized to LRU) compared to estimated development time of replacement policies for the single-core prefetch-enabled ChampSim configuration on memory-intensive SPEC06 workloads. Slope (grey, italics) denotes percentage point improvement per day.
  • Figure 5: Ablation study measuring improvement in suite-level geomean IPC speedup with each new technique that composes Policy31 for the single-core prefetch-enabled ChampSim configuration running SPEC06 memory intensive workloads.
  • ...and 4 more figures