# AscendOptimizer: Episodic Agent for Ascend NPU Operator Optimization

arXiv: https://arxiv.org/abs/2603.23566 · HTML: https://mlsanity.com/paper/2603.23566/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 38 of its 52 references that are in the map.
- First posted on arXiv 24 Mar 2026.

## Builds on (most cited first, 30 of 38)

- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [Voyager: An Open-Ended Embodied Agent with Large Language Models](https://mlsanity.com/paper/2305.16291/) (2023)
- [Hindsight Experience Replay](https://mlsanity.com/paper/1707.01495/) (2017)
- [FlashAttention-3: Fast and Accurate Attention with Asynchrony and Low-precision](https://mlsanity.com/paper/2407.08608/) (2024)
- [The Stack: 3 TB of permissively licensed source code](https://mlsanity.com/paper/2211.15533/) (2022)
- [KernelBench: Can LLMs Write Efficient GPU Kernels?](https://mlsanity.com/paper/2502.10517/) (2025)
- [Ansor: Generating High-Performance Tensor Programs for Deep Learning](https://mlsanity.com/paper/2006.06762/) (2020)
- [Kevin: Multi-Turn RL for Generating CUDA Kernels](https://mlsanity.com/paper/2507.11948/) (2025)
- [TritonBench: Benchmarking Large Language Model Capabilities for Generating Triton Operators](https://mlsanity.com/paper/2502.14752/) (2025)
- [CUDA-L1: Improving CUDA Optimization via Contrastive Reinforcement Learning](https://mlsanity.com/paper/2507.14111/) (2025)
- [ReWiND: Language-Guided Rewards Teach Robot Policies without New Demonstrations](https://mlsanity.com/paper/2505.10911/) (2025)
- [Self-Supervised Bug Detection and Repair](https://mlsanity.com/paper/2105.12787/) (2021)
- [Tiramisu: A Polyhedral Compiler for Expressing Fast and Portable Code](https://mlsanity.com/paper/1804.10694/) (2018)
- [Astra: A Multi-Agent System for GPU Kernel Performance Optimization](https://mlsanity.com/paper/2509.07506/) (2025)
- [Seed-Coder: Let the Code Model Curate Data for Itself](https://mlsanity.com/paper/2506.03524/) (2025)
- [The Deep Learning Compiler: A Comprehensive Survey](https://mlsanity.com/paper/2002.03794/) (2020)
- [CudaForge: An Agent Framework with Hardware Feedback for CUDA Kernel Optimization](https://mlsanity.com/paper/2511.01884/) (2025)
- [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)
- [AutoTriton: Automatic Triton Programming with Reinforcement Learning in LLMs](https://mlsanity.com/paper/2507.05687/) (2025)
- [MultiKernelBench: A Multi-Platform Benchmark for Kernel Generation](https://mlsanity.com/paper/2507.17773/) (2025)
- [SelfAPR: Self-supervised Program Repair with Test Execution Diagnostics](https://mlsanity.com/paper/2203.12755/) (2022)
- [Geak: Introducing Triton Kernel AI Agent & Evaluation Benchmarks](https://mlsanity.com/paper/2507.23194/) (2025)
- [Unveiling Memorization in Code Models](https://mlsanity.com/paper/2308.09932/) (2023)
- [STARK: Strategic Team of Agents for Refining Kernels](https://mlsanity.com/paper/2510.16996/) (2025)
- [TritonRL: Training LLMs to Think and Code Triton Without Cheating](https://mlsanity.com/paper/2510.17891/) (2025)
- [Daydream: Accurately Estimating the Efficacy of Optimizations for DNN Training](https://mlsanity.com/paper/2006.03318/) (2020)
- [GPU Kernel Scientist: An LLM-Driven Framework for Iterative Kernel Optimization](https://mlsanity.com/paper/2506.20807/) (2025)

## Built on by (most cited first, 3 of 3)

- [SLAI T-Rex: Full-Parameter Post-training of the DeepSeek-V4 Family on Ascend SuperPOD](https://mlsanity.com/paper/2607.20145/) (2026)
- [CANN Bench: Benchmarking Agent Generated Kernels against Real NPU and Algorithmic Limits](https://mlsanity.com/paper/2607.20518/) (2026)
- [From Experience to Expertise: Adoption-Aware Memory Learning for Data-Scarce NPU Kernel Synthesis](https://mlsanity.com/paper/2609.35568/) (2026)

## Abstract (arXiv)

Optimizing AscendC (Ascend C) operators for Ascend NPUs is difficult for two reasons. First, unlike CUDA, the ecosystem offers few public kernels to learn from. Second, performance depends on a coupled two-part implementation: a host-side tiling program that controls data movement and a kernel program that schedules and pipelines computation. We present AscendOptimizer, an episodic agent that builds missing optimization knowledge from execution itself. For kernel optimization, AscendOptimizer rewinds strong implementations by removing optimizations in a controlled way, then keeps the changes whose removal measurably hurts performance as reusable experience for later rewriting. For host-side optimization, it runs profiling-in-the-loop evolutionary search to find valid, fast tiling and data-movement configurations directly from hardware feedback. This combination lets the agent improve kernel structure and host-side scheduling together. On a benchmark of 101 real AscendC operators, AscendOptimizer achieves a 1.21x geometric-mean speedup over the open-source baseline, and 53.47% of operators run faster than their references. Given a same budget of evaluations per operator, AscendOptimizer consistently outperforms Best-of-N sampling and OpenEvolve in terms of geometric mean speedup, fast_p tail speedup ratios, and overall optimization progress across varying budgets.