# Axon: A Synthesizing Superoptimizer for Tensor Programs

arXiv: https://arxiv.org/abs/2606.26344 · HTML: https://mlsanity.com/paper/2606.26344/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 12 of its 55 references that are in the map.
- First posted on arXiv 24 Jun 2026.

## Builds on (most cited first, 12 of 12)

- [Root Mean Square Layer Normalization](https://mlsanity.com/paper/1910.07467/) (2019)
- [The Falcon Series of Open Language Models](https://mlsanity.com/paper/2311.16867/) (2023)
- [Ansor: Generating High-Performance Tensor Programs for Deep Learning](https://mlsanity.com/paper/2006.06762/) (2020)
- [Learning to Optimize Tensor Programs](https://mlsanity.com/paper/1805.08166/) (2018)
- [TileLang: A Composable Tiled Programming Model for AI Systems](https://mlsanity.com/paper/2504.17577/) (2025)
- [Mirage: A Multi-Level Superoptimizer for Tensor Programs](https://mlsanity.com/paper/2405.05751/) (2024)
- [TensorIR: An Abstraction for Automatic Tensorized Program Optimization](https://mlsanity.com/paper/2207.04296/) (2022)
- [Hidet: Task-Mapping Programming Paradigm for Deep Learning Tensor Programs](https://mlsanity.com/paper/2210.09603/) (2022)
- [Equality Saturation for Tensor Graph Superoptimization](https://mlsanity.com/paper/2101.01332/) (2021)
- [TensorRight: Automated Verification of Tensor Graph Rewrites](https://mlsanity.com/paper/2511.17838/) (2025)
- [Pushing Tensor Accelerators Beyond MatMul in a User-Schedulable Language](https://mlsanity.com/paper/2512.02371/) (2025)
- [Nice to Meet You: Synthesizing Practical MLIR Abstract Transformers](https://mlsanity.com/paper/2512.06442/) (2025)

## Abstract (arXiv)

Writing high performance kernels for AI accelerators requires deep expertise in tiling, instruction selection, data layout, and operator fusion placing a significant burden on programmers. In this paper, we focus on tile based AI accelerator programs and present Axon, a synthesizing superoptimizer for tensor programs: it uses program synthesis to automatically generate target instructions from semantics specifications, and explores semantically equivalent program variants to select the best performing kernel empirically. Axon discovers algebraic transformations by propagating operators through computation graphs and uses SMT over unbounded tensors to guarantee that all transformations preserve semantics without requiring hand crafted rewrite rules. It then lowers tensor operations to target ISA instructions, explores tiling configurations constrained by hardware descriptions, and fuses operators and instructions to minimize memory traffic.