# PRAGMA: A Profiling-Reasoned Multi-Agent Framework for Automatic Kernel Optimization

arXiv: https://arxiv.org/abs/2511.06345 · HTML: https://mlsanity.com/paper/2511.06345/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning.
- Builds on 9 of its 22 references that are in the map.
- First posted on arXiv 9 Nov 2025.

## Placed in

- LLM reasoning / [Reasoning for agents, planning and code](https://mlsanity.com/lr/reasoning-for-agents-planning-and-code/)
- LLM agents and tool use / [Software engineering coding agents](https://mlsanity.com/agents/software-engineering-coding-agents/)

## Builds on (most cited first, 9 of 9)

- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [ChatDev: Communicative Agents for Software Development](https://mlsanity.com/paper/2307.07924/) (2023)
- [KernelBench: Can LLMs Write Efficient GPU Kernels?](https://mlsanity.com/paper/2502.10517/) (2025)
- [Astra: A Multi-Agent System for GPU Kernel Performance Optimization](https://mlsanity.com/paper/2509.07506/) (2025)
- [SwizzlePerf: Hardware-Aware LLMs for GPU Kernel Performance Optimization](https://mlsanity.com/paper/2508.20258/) (2025)
- [MARCO: Multi-Agent Code Optimization with Real-Time Knowledge Integration for High-Performance Computing](https://mlsanity.com/paper/2505.03906/) (2025)
- [QiMeng-TensorOp: Automatically Generating High-Performance Tensor Operators with Hardware Primitives](https://mlsanity.com/paper/2505.06302/) (2025)
- [QiMeng-Attention: SOTA Attention Operator is generated by SOTA Attention Algorithm](https://mlsanity.com/paper/2506.12355/) (2025)

## Built on by (most cited first, 11 of 11)

- [KernelFoundry: Hardware-aware evolutionary GPU kernel optimization](https://mlsanity.com/paper/2603.12440/) (2026)
- [Towards Automated Kernel Generation in the Era of LLMs](https://mlsanity.com/paper/2601.15727/) (2026)
- [KernelSkill: A Multi-Agent Framework for GPU Kernel Optimization](https://mlsanity.com/paper/2603.10085/) (2026)
- [AscendCraft: Automatic Ascend NPU Kernel Generation via DSL-Guided Transcompilation](https://mlsanity.com/paper/2601.22760/) (2026)
- [Kernel-Smith: A Unified Recipe for Evolutionary Kernel Optimization](https://mlsanity.com/paper/2603.28342/) (2026)
- [AscendOptimizer: Episodic Agent for Ascend NPU Operator Optimization](https://mlsanity.com/paper/2603.23566/) (2026)
- [PerfAgent: Profiler-Guided Iterative Refinement for Repository-Level Code Optimization](https://mlsanity.com/paper/2607.19653/) (2026)
- [From Tokens to Regions: CUDA-Sensitive Instruction Tuning for GPU Kernel Generation](https://mlsanity.com/paper/2606.16231/) (2026)
- [ParEVO: Synthesizing Code for Irregular Data: High-Performance Parallelism through Agentic Evolution](https://mlsanity.com/paper/2603.02510/) (2026)
- [Towards Feedback-to-Plan Decisions for Self-Evolving LLM Agents in CUDA Kernel Generation](https://mlsanity.com/paper/2605.26720/) (2026)
- [Compiler-Grounded Hierarchical Diagnosis for LLM-Based Triton Kernel Optimization](https://mlsanity.com/paper/2607.23089/) (2026)

## Abstract (arXiv)

Designing high-performance kernels requires expert-level tuning and a deep understanding of hardware characteristics. Recent advances in large language models (LLMs) have enabled automated kernel generation, yet most existing systems rely solely on correctness or execution time feedback, lacking the ability to reason about low-level performance bottlenecks. In this paper, we introduce PRAGMA, a profile-guided AI kernel generation framework that integrates execution feedback and fine-grained hardware profiling into the reasoning loop. PRAGMA enables LLMs to identify performance bottlenecks, preserve historical best versions, and iteratively refine code quality. We evaluate PRAGMA on KernelBench, covering GPU and CPU backends. Results show that PRAGMA consistently outperforms baseline AIKG without profiling enabled and achieves 2.81× and 2.30× averaged speedups against Torch on CPU and GPU platforms, respectively.