Pith Number

pith:Z4VOBKGX

pith:2025:Z4VOBKGXGFZWTTJPII5LRL5G32

not attested not anchored not stored refs resolved

Optimizing PyTorch Inference with LLM-Based Multi-Agent Systems

Costin Iancu, Kirill Nagaitsev, Luka Grbcic, Samuel Williams

Multi-agent LLM systems optimize PyTorch code for 2.88x faster inference than eager execution on H100 GPUs.

arxiv:2511.16964 v2 · 2025-11-21 · cs.MA · cs.AI · cs.DC

Open paper page JSON Open Graph Bundle Merged state What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{Z4VOBKGXGFZWTTJPII5LRL5G32}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

The best implementation achieves an average 2.88x speedup over PyTorch Eager (1.85x over torch.compile) on an H100 GPU across diverse tasks in KernelBench.

C2weakest assumption

That LLM-based multi-agent systems can reliably generate correct, bug-free optimized code at scale and that results on the KernelBench suite generalize to production inference workloads.

C3one line summary

An exploit-heavy multi-agent LLM system with error-fixing agents delivers 2.88x average speedup over PyTorch Eager and 1.85x over torch.compile on H100 GPUs across KernelBench tasks.

References

5 extracted · 5 resolved · 1 Pith anchors

[1] Accessed: 2025-10-17 2025 · doi:10.1145/3703412.3703416

[2] AlphaEvolve: A coding agent for scientific and algorithmic discovery 2025 · arXiv:2506.13131

[3] Anjiang Wei, Tianran Sun, Yogesh Seenichamy, Hang Song, Anne Ouyang, Azalia Mirhoseini, Ke Wang, and Alex Aiken 2019

[4] Zhang, Z., Bajaj, A 2025

[5] ISBN 9798331314385 2025

Formal links

1 machine-checked theorem link

Receipt and verification

First computed	2026-05-17T23:39:17.078540Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

cf2ae0a8d7317369cd2f423ab8afa6de9bc28f425416eb1a8ce9b3d4dd32297b

Aliases

arxiv: 2511.16964 · arxiv_version: 2511.16964v2 · doi: 10.48550/arxiv.2511.16964 · pith_short_12: Z4VOBKGXGFZW · pith_short_16: Z4VOBKGXGFZWTTJP · pith_short_8: Z4VOBKGX

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/Z4VOBKGXGFZWTTJPII5LRL5G32 \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: cf2ae0a8d7317369cd2f423ab8afa6de9bc28f425416eb1a8ce9b3d4dd32297b

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "1ff1dec44e5b433b61e42ec1111f63dac46c3621afeb64856727183cf86e0a1f",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.DC"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.MA",
    "submitted_at": "2025-11-21T05:37:38Z",
    "title_canon_sha256": "08f42973845b6848287e8643acf8abaae8107c35a8a8905d5b8eca6c607c6e44"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2511.16964",
    "kind": "arxiv",
    "version": 2
  }
}