{"slug": "schedules-are-solvable-symbols-tuning-free-compilation-of-tile-programs", "title": "Schedules Are Solvable Symbols: Tuning-Free Compilation of Tile Programs", "summary": "Researchers submitted a paper on 24 Sep 2026 presenting Loom, a tuning-free symbolic compiler framework for tile-based SPMD programs on spatial dataflow architectures. Loom formulates one CP-SAT problem per schedule candidate, jointly solving inter-core dataflow, intra-core asynchronous scheduling, and block sizes at compile time, and on Tenstorrent's Wormhole and Blackhole generations it matches or exceeds the vendor-optimized TTNN library on GEMM, Flash Attention, and Flash Decode without per-shape profiling or profile-based platform-specific schedule tuning.", "body_md": "# Computer Science > Programming Languages\n\n  [Submitted on 24 Sep 2026]\n\n# Title:Schedules Are Solvable Symbols: Tuning-Free Compilation of Tile Programs on Dataflow Architectures\n\n[View PDF](https://arxiv.org/pdf/2609.29219)\n\n[HTML (experimental)](https://arxiv.org/html/2609.29219v1)\n\nAbstract:Modern AI and HPC accelerators increasingly expose dataflow features: software-visible mechanisms for data movement and overlap, such as inter-core communication through the on-chip network and intra-core asynchronous pipelining. These features shift scheduling responsibility from hardware to the compiler, and because placement, movement, and synchronization become software-visible, they also make the performance of static schedules predictable. Yet high performance on such hardware still relies on vendor-engineered kernel libraries or profile-based auto-tuning, whose embedded expert knowledge transfers poorly across architectures and algorithms.\n\nWe present Loom, a tuning-free symbolic compiler framework for tile-based SPMD programs on spatial dataflow architectures. The central idea is to treat tile-based SPMD compilation as a hardware-explicit static optimization problem. Loom enumerates discrete spatial-mapping and communication candidates while keeping value parameters, such as tiling factors and pipeline knobs, symbolic within each candidate.\n\nFrom an explicit hardware description, it derives symbolic legality constraints and latency expressions, formulates one CP-SAT problem per schedule candidate, and jointly solves inter-core dataflow, intra-core asynchronous scheduling, and block sizes at compile time.\n\nOn two Tenstorrent generations, Wormhole and Blackhole, Loom matches or exceeds the vendor-optimized TTNN library on GEMM, Flash Attention, and Flash Decode, out of the box and without per-shape profiling or profile-based platform-specific schedule tuning. These results suggest that hardware-derived symbolic compilation provides a retargetable alternative to profiling-based tuning for spatial dataflow architectures while remaining interpretable by keeping optimization decisions traceable to source-level symbols.\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/))\n# 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))\n# 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))\n# 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/schedules-are-solvable-symbols-tuning-free-compilation-of-tile-programs", "canonical_source": "https://arxiv.org/abs/2609.29219", "published_at": "2026-09-29 18:02:17+00:00", "updated_at": "2026-09-29 18:17:59.607321+00:00", "lang": "en", "topics": ["ai-infrastructure", "ai-chips", "mlops", "ai-research", "developer-tools"], "entities": ["Loom", "Tenstorrent", "Wormhole", "Blackhole", "TTNN", "arXiv"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/schedules-are-solvable-symbols-tuning-free-compilation-of-tile-programs", "markdown": "https://wpnews.pro/news/schedules-are-solvable-symbols-tuning-free-compilation-of-tile-programs.md", "text": "https://wpnews.pro/news/schedules-are-solvable-symbols-tuning-free-compilation-of-tile-programs.txt", "jsonld": "https://wpnews.pro/news/schedules-are-solvable-symbols-tuning-free-compilation-of-tile-programs.jsonld"}}