{"slug": "a-contract-grade-verifier-for-llm-generated-gpu-kernels", "title": "A Contract-Grade Verifier for LLM-Generated GPU Kernels", "summary": "A new contract-grade verifier with twelve adversarial gates found that 39.5% of 2,638 machine-generated GPU kernels accepted as correct by a public system's harness are broken beyond any tolerance argument, and 62.1% carry at least one violation. The verifier, described in a paper submitted to arXiv on 13 Aug 2026, also introduces the first native Blackwell tcgen05 training backward for the gated-linear-recurrence (GDN) family, validated against a double-precision oracle. The findings suggest that the correctness signal behind reported progress in kernel generation is far weaker than the numbers suggest.", "body_md": "# Computer Science > Machine Learning\n\n[Submitted on 13 Aug 2026]\n\n# Title:A Contract-Grade Verifier for LLM-Generated GPU Kernels, and a Native Blackwell Backward for the Gated-Linear-Recurrence Family\n\n[View PDF](/pdf/2608.12700)\n\n[HTML (experimental)](https://arxiv.org/html/2608.12700v1)\n\nAbstract:Systems that generate GPU kernels with language models report high correctness rates. Those rates come from a single loose test: run the kernel on a few random inputs at one fixed shape and accept it if the output is close to a reference. A kernel can pass that test and still be silently wrong. It can return an ordinary number where the true answer is a NaN or an infinity, differ from run to run, break when the shape changes, or accumulate in fp16 where the reference keeps an fp32 total. We build the instrument that checks correctness properly: a contract-grade verifier of twelve adversarial gates, each a property a correct kernel must satisfy, several of them tolerance-free, so no choice of threshold can explain a failure away. Aimed outward, the verifier audits 2,638 machine-generated kernels that a public system's own harness had already accepted as correct. It finds 39.5% broken beyond any tolerance argument and 62.1% carrying at least one violation. The field's standard test accepts 1,487 kernels the verifier rejects, against only 14 the other way. We defend the finding four independent ways: a 7/7 positive control, a threshold-calibration sweep, 98.5% agreement with the reference benchmark's own correctness code, and a stratified hand-audit. Aimed inward, the verifier judges a kernel of our own: the first native Blackwell tcgen05 training backward for the gated-linear-recurrence (GDN) family, including the reverse-state stage the field still runs on a fallback. We establish its correctness independently, against a double-precision oracle, and train five family members through it. The correctness signal behind reported progress in kernel generation is far weaker than the numbers suggest, and a set of tolerance-free contracts would close most of the gap.\n\n### Current browse context:\n\ncs.LG\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer\n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers\n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps\n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations\n\n*(*[What are Smart Citations?](https://www.scite.ai/))# Code, Data and Media Associated with this Article\n\nalphaXiv\n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers\n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub\n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub\n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face\n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast\n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower\n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender\n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))\nIArxiv Recommender\n\n*(*[What is IArxiv?](https://iarxiv.org/about))# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth 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.\n\nHave an idea for a project that will add value for arXiv's community? [ Learn more about arXivLabs](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/a-contract-grade-verifier-for-llm-generated-gpu-kernels", "canonical_source": "https://arxiv.org/abs/2608.12700", "published_at": "2026-08-14 16:57:14+00:00", "updated_at": "2026-08-14 17:11:51.561343+00:00", "lang": "en", "topics": ["machine-learning", "artificial-intelligence", "ai-research"], "entities": ["arXiv", "Blackwell", "GDN"], "alternates": {"html": "https://wpnews.pro/news/a-contract-grade-verifier-for-llm-generated-gpu-kernels", "markdown": "https://wpnews.pro/news/a-contract-grade-verifier-for-llm-generated-gpu-kernels.md", "text": "https://wpnews.pro/news/a-contract-grade-verifier-for-llm-generated-gpu-kernels.txt", "jsonld": "https://wpnews.pro/news/a-contract-grade-verifier-for-llm-generated-gpu-kernels.jsonld"}}