Skip to content
AI News HubLIVE
Original source2 min read

PerfReasoning: How Well Do LLMs Reason on Hardware Performance?

Summary

A new benchmark, PerfReasoning, evaluates LLMs both as direct performance reasoners and as generators of analytical performance-model code. Top closed-source models exceed 90% on reasoning Q&A and the best open-weight model reaches 82.4%, but model construction is far harder: apart from GPT-5.6 Sol, configurations average below 15% pass rates with high run-to-run variance.

SourcearXiv AIAuthor: Dan Zhao, Karthikeyan Sankaralingam, Christos Kozyrakis, Qijing Huang
PerfReasoning: How Well Do LLMs Reason on Hardware Performance?
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 3 Sep 2026]

Title:PerfReasoning: How Well Do LLMs Reason on Hardware Performance?

View a PDF of the paper titled PerfReasoning: How Well Do LLMs Reason on Hardware Performance?, by Dan Zhao and 3 other authors

View PDF HTML (experimental)

Abstract:Performance modeling is central to hardware design and software optimization, yet constructing these models requires structured reasoning about computation, data reuse, storage, and movement. We introduce PerfReasoning, a benchmark that evaluates LLMs both as direct performance reasoners and as generators of analytical performance-model code. Given workload, architecture, and mapping specifications, models compare mappings and predict off-chip traffic and buffer requirements. The strongest closed-source models exceed 90% on reasoning-based Q&A, and the best open-weight model reaches 82.4%. However, model construction is substantially harder: while GPT-5.6 Sol exceeds 80% pass rate, all other model configurations average below 15% and vary markedly across runs. Task-specific RL raises a 4B model's mapping-reasoning accuracy by 15.7 points, whereas feedback-free multi-round self-revision prompting is not reliably effective. PerfReasoning exposes the gap between plausible architectural reasoning and reliable performance-model construction. We will publicly release the benchmark to support reproducible evaluation and track future progress.

Subjects:

Artificial Intelligence (cs.AI); Performance (cs.PF)

Cite as: arXiv:2609.04476 [cs.AI]

(or arXiv:2609.04476v1 [cs.AI] for this version)

https://doi.org/10.48550/arXiv.2609.04476

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Qijing Huang [view email] [v1] Thu, 3 Sep 2026 21:00:06 UTC (37 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled PerfReasoning: How Well Do LLMs Reason on Hardware Performance?, by Dan Zhao and 3 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.AI

new | recent | 2026-09

Change to browse by:

cs cs.PF

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • PerfReasoning tests LLMs on direct performance reasoning and on generating analytical performance-model code.
  • The strongest closed-source models exceed 90% on reasoning Q&A; the best open-weight model reaches 82.4%.
  • Model construction is much harder: GPT-5.6 Sol exceeds 80% pass rate while all other configurations average below 15% with marked variance across runs.
  • Task-specific RL raises a 4B model's mapping-reasoning accuracy by 15.7 points, whereas feedback-free multi-round self-revision prompting is not reliably effective.

Highlights and analysis are generated automatically and may contain errors. Check the original source.