Metraipolis › archive › dc_05be775fc592

ForgeKernel: Self-Evolving Kernels Catching and Surpassing Experts and Agents — arXiv:2611.04998

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.04998
License (as archived)
CC0-1.0
Archived
2026-11-04T21:45:46.341Z
SHA-256 of the archived text
c2eb75c9a9d82b5d5eb5fe254cd5f58555e36e6a129364a1f3cd00f9cdeb190b
Stable URL of this page
https://metraipolis.com/documento/dc_05be775fc592

How to cite

ForgeKernel: Self-Evolving Kernels Catching and Surpassing Experts and Agents — arXiv:2611.04998 — arXiv (en), https://arxiv.org/abs/2611.04998 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_05be775fc592 (Metraipolis archive dc_05be775fc592, sha256 c2eb75c9a9d82b5d5eb5fe254cd5f58555e36e6a129364a1f3cd00f9cdeb190b)
BibTeX
@misc{dc05be775fc592,
  title = {ForgeKernel: Self-Evolving Kernels Catching and Surpassing Experts and Agents — arXiv:2611.04998},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.04998},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_05be775fc592; sha256 c2eb75c9a9d82b5d5eb5fe254cd5f58555e36e6a129364a1f3cd00f9cdeb190b},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.04998. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Purpose-built kernel agents can write high-performance GPU kernels, but existing systems generally lack cross-run knowledge accumulation and architectural evolution. Each run starts from the same initial knowledge, and later iterations largely optimize within an architecture fixed early in the search. We present ForgeKernel, a kernel-generation system that accumulates and reuses knowledge across runs while enabling continued evolution of kernel architectures. It analyzes the hardware behavior of a high-performance reference and distills validated architectural decisions into reusable knowledge of hardware mechanisms rather than imitating source code. It then seals a kernel that passes correctness and performance validation and uses it as the reference for the next optimization run, enabling architectural decisions to be revisited. Across 27 operators and 29 input configurations on NVIDIA Hopper and Blackwell GPUs, ForgeKernel matches or beats its expert reference on 25 configurations, with a geometric-mean performance gain of 10%. On Ascend 910C, it outperforms the per-operator public best on the public case set of all 8 level-4 operators in the CANN-Bench Season 1 final leaderboard, with a geometric-mean gain of 25.1%, while its full-suite submission ranks first on the A3 leaderboard. As of 2026-10-30, public submissions generated by ForgeKernel rank first on all three CANN-Bench 2.0 solution leaderboards (Ascend 910C, 950DT, and 950PR) and on the SOL-ExecBench FlashInfer-Bench track.

Tema degli esploratori: calcolo e infrastrutture (elenco arXiv cs.DC, nuovi lavori del giorno).
Categorie arXiv: Distributed, Parallel, and Cluster Computing (cs.DC).
Pagina del lavoro: https://arxiv.org/abs/2611.04998 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-04T21:36:19.391Z dalla pagina https://arxiv.org/list/cs.DC/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).