This work presents MaxKernel, a multi-agent system that implements three distinct paradigms for TPU kernel development: a Human-in-the-Loop (HITL) agent for collaborative, step-by-step design; an Autonomous (Auto) agent that executes a fully automated, metric/trace-driven optimization loop; and a Graph-Based Autonomous Search that scales the Auto agent for global exploration of the design space.
Abstract
Designing and authoring high-performance custom kernels for accelerators is a complex task that requires deep hardware-level expertise. Large Language Models (LLM) can be leveraged together with real-time compiler feedback to build agentic systems for kernel generation. In this work, we present MaxKernel, a multi-agent system that implements three distinct paradigms for TPU kernel development: (1) a Human-in-the-Loop (HITL) agent for collaborative, step-by-step design; (2) an Autonomous (Auto) agent that executes a fully automated, metric/trace-driven optimization loop; and (3) a Graph-Based Autonomous Search that scales the Auto agent for global exploration of the design space. All three paradigms leverage a shared pool of specialized sub-agents to handle planning, implementation, self-debugging, testing, and hardware profiling. We evaluate MaxKernel on JaxBench, a comprehensive suite of 50 diverse kernel tasks for TPUs, alongside complex, real-world workloads from state-of-the-art open-source models. We demonstrate that MaxKernel consistently generates highly optimized implementations, matching expert hand-tuned baselines and delivering significant performance across the benchmark. Our agent is open-sourced and available https://github.com/AI-Hypercomputer/accelerator-agents/tree/main/MaxKernel.
LLM4LLM is introduced, a deployment-aware closed-loop optimization framework that starts from a target inference script, extracts phase-aware optimization tasks, searches with an experience-guided episodic agent, and accepts patches through in-model validation.
Hui Zeng, Pengfei Yang, Yanxin Chen et al.· 0 citations
Argus is introduced, an agentic LLM-based profiler that produces instrumentation code and autonomously reasons over potential OS-level bottlenecks and aims to identify a specific kernel code path rather than stop at a subsystem-level diagnosis.
CAKE, a compiler-agent co-design in which agents author CAKE IR, a typed, hardware-explicit schedule representation, exposes warp roles, memory movement, synchronization, and pipelines while supporting verification, cost modeling, and localized diagnostics.
Zihao Ye, Yingyi Huang, H. Jin et al.· 4 citations
The chat-to-agentic gap in current inference benchmarks is quantified and per-kernel GPU resource utilisation via roofline analysis is characterised via roofline analysis.
Cheuk H. Lau, Ze-Yu Cao, K. Yin et al.· arXiv.org· 2 citations
We introduce AMDKernelVault, an open HIP and Triton kernel corpus and training framework for recent AMD CDNA GPUs. Existing LLM-based kernel agents are largely CUDA/NVIDIA-centric and often depend on repeated frontier-LLM calls for generation, reflection, and optimization. To address this gap, we develop HIPKernelGen a...
Ji Liu, S. Majumder, Yi-Qing Huang et al.· 0 citations
KernelArc, a multi-agent framework for autonomous GPU kernel optimization across heterogeneous workloads, can broaden exploration and reach stronger incumbents within a fixed candidate budget, while the value of individual coordination features depends on the kernel and optimization stage.
Joyjit Kundu, Ben Stoffelen, Kaili Wang et al.· 1 citation