---
title: "Ask, Don't Equip: Question-Driven Coding Agents for Deep Kernel Optimization"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2611.03517v1"
permalink: "https://feed.astralor.com/items/itm_1e5a4855bc738d69c2898473"
published: "2026-11-02T17:51:58.000Z"
score: 61
channel: "tech"
tags: ["LLM","AI Agent","RAG","Ask","Don","Equip","Question-Driven Coding Agents","Deep Kernel Optimization"]
---

# Ask, Don't Equip: Question-Driven Coding Agents for Deep Kernel Optimization

## At a glance

LLM agents for GPU kernel optimization sit at two ends of a structure-freedom tradeoff. Scripted agents fix what the model does at each step, via a learned policy, an optimization menu, or fixed roles, and are sample-ef…

## Key points

- LLM agents for GPU kernel optimization sit at two ends of a structure-freedom tradeoff.
- Scripted agents fix what the model does at each step, via a learned policy, an optimization menu, or fixed roles, and are sample-efficient where the recipe applies and stall where it does not.
- Open-ended agents let a self-directed model mutate a population of kernels and pay for the freedom in trials: thousands of evaluations, repeated edits, and wrong candidates.

## Full text

LLM agents for GPU kernel optimization sit at two ends of a structure-freedom tradeoff. Scripted agents fix what the model does at each step, via a learned policy, an optimization menu, or fixed roles, and are sample-efficient where the recipe applies and stall where it does not. Open-ended agents let a self-directed model mutate a population of kernels and pay for the freedom in trials: thousands of evaluations, repeated edits, and wrong candidates. Both respond to hard targets by equipping the agent with more: parsed profiler output, disassembly analysis, retrieval over documentation and libraries, and stores of lessons. We argue that both lack direction, a hypothesis about where the remaining gap is, and that a short question asked at the right moment supplies it without narrowing what the agent may do next. We present KernelSeek, a question-driven agent that makes an existing CUDA kernel faster with no operator-specific knowledge and no faster implementation to imitate: a general-purpose coding agent, a minimal harness that enforces a verifiable goal, and questions at decision points, first from a human, then from a bank the agent applies to itself. On the same task, evaluator, hardware, and week, a tool-rich autonomous agent built in the prevailing style plateaued at 0.50x FlashAttention-3 after 133 model-driven steps; KernelSeek reached 1.13x. It then produced dependency-free FlashAttention forward kernels for Hopper and Blackwell, single CUDA files with no CUTLASS, CuTe, cuBLAS, or cuDNN, correct on all 182 scenarios per architecture and 1.16x FlashAttention-3 and 1.46x FlashAttention-4 (geometric mean). Across 162 human messages and 11.5K agent tool calls, every large gain followed a question that moved the agent to a different layer of the system or rejected a claimed limit, never one that supplied information. We release the kernels, harness, question bank, and all traces.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_1e5a4855bc738d69c2898473