# PerfDojo: Automated ML Library Generation for Heterogeneous Architectures

arXiv: https://arxiv.org/abs/2511.03586 · HTML: https://mlsanity.com/paper/2511.03586/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 25 of its 57 references that are in the map.
- First posted on arXiv 5 Nov 2025.

## Builds on (most cited first, 25 of 25)

- [Deep Reinforcement Learning with Double Q-learning](https://mlsanity.com/paper/1509.06461/) (2015)
- [Prioritized Experience Replay](https://mlsanity.com/paper/1511.05952/) (2015)
- [Graph of Thoughts: Solving Elaborate Problems with Large Language Models](https://mlsanity.com/paper/2308.09687/) (2023)
- [Dueling Network Architectures for Deep Reinforcement Learning](https://mlsanity.com/paper/1511.06581/) (2015)
- [FlashAttention-3: Fast and Accurate Attention with Asynchrony and Low-precision](https://mlsanity.com/paper/2407.08608/) (2024)
- [Noisy Networks for Exploration](https://mlsanity.com/paper/1706.10295/) (2017)
- [Reinforcement learning](https://mlsanity.com/paper/2405.10369/) (2024)
- [TPU v4: An Optically Reconfigurable Supercomputer for Machine Learning with Hardware Support for Embeddings](https://mlsanity.com/paper/2304.01433/) (2023)
- [Ansor: Generating High-Performance Tensor Programs for Deep Learning](https://mlsanity.com/paper/2006.06762/) (2020)
- [Tiramisu: A Polyhedral Compiler for Expressing Fast and Portable Code](https://mlsanity.com/paper/1804.10694/) (2018)
- [MARLIN: Mixed-Precision Auto-Regressive Parallel Inference on Large Language Models](https://mlsanity.com/paper/2408.11743/) (2024)
- [CompilerGym: Robust, Performant Compiler Optimization Environments for AI Research](https://mlsanity.com/paper/2109.08267/) (2021)
- [Machine Learning Model Sizes and the Parameter Gap](https://mlsanity.com/paper/2207.02852/) (2022)
- [TensorIR: An Abstraction for Automatic Tensorized Program Optimization](https://mlsanity.com/paper/2207.04296/) (2022)
- [Tensor Program Optimization with Probabilistic Programs](https://mlsanity.com/paper/2205.13603/) (2022)
- [Stateful Dataflow Multigraphs: A Data-Centric Model for Performance Portability on Heterogeneous Architectures](https://mlsanity.com/paper/1902.10345/) (2019)
- [Bolt: Bridging the Gap between Auto-tuners and Hardware-native Performance](https://mlsanity.com/paper/2110.15238/) (2021)
- [Snitch: A tiny Pseudo Dual-Issue Processor for Area and Energy Efficient Execution of Floating-Point Intensive Workloads](https://mlsanity.com/paper/2002.10143/) (2020)
- [Maximum Reward Formulation In Reinforcement Learning](https://mlsanity.com/paper/2010.03744/) (2020)
- [Stream Semantic Registers: A Lightweight RISC-V ISA Extension Achieving Full Compute Utilization in Single-Issue Cores](https://mlsanity.com/paper/1911.08356/) (2019)
- [Ara: A 1 GHz+ Scalable and Energy-Efficient RISC-V Vector Processor with Multi-Precision Floating Point Support in 22 nm FD-SOI](https://mlsanity.com/paper/1906.00478/) (2019)
- [Bring Your Own Codegen to Deep Learning Compiler](https://mlsanity.com/paper/2105.03215/) (2021)
- [Exo 2: Growing a Scheduling Language](https://mlsanity.com/paper/2411.07211/) (2024)
- [ISA Mapper: A Compute and Hardware Agnostic Deep Learning Compiler](https://mlsanity.com/paper/1810.09958/) (2018)
- [FuzzyFlow: Leveraging Dataflow To Find and Squash Program Optimization Bugs](https://mlsanity.com/paper/2306.16178/) (2023)

## Abstract (arXiv)

The increasing complexity of machine learning models and the proliferation of diverse hardware architectures (CPUs, GPUs, accelerators) make achieving optimal performance a significant challenge. Heterogeneity in instruction sets, specialized kernel requirements for different data types and model features (e.g., sparsity, quantization), and architecture-specific optimizations complicate performance tuning. Manual optimization is resource-intensive, while existing automatic approaches often rely on complex hardware-specific heuristics and uninterpretable intermediate representations, hindering performance portability. We introduce PerfLLM, a novel automatic optimization methodology leveraging Large Language Models (LLMs) and Reinforcement Learning (RL). Central to this is PerfDojo, an environment framing optimization as an RL game using a human-readable, mathematically-inspired code representation that guarantees semantic validity through transformations. This allows effective optimization without prior hardware knowledge, facilitating both human analysis and RL agent training. We demonstrate PerfLLM's ability to achieve significant performance gains across diverse CPU (x86, Arm, RISC-V) and GPU architectures.