FABRICA: Agentic CUDA-to-CSL Translation and Optimization for Wafer-Scale Systems

2026-08-25

AI Summary

研究提出FABRICA框架，通过结合目标知识、执行、故障修复和优化，解决GPU内核在不同架构间移植的问题，显著提高了CUDA到CSL转换的成功率和性能。

Abstract

Porting GPU kernels across architectures requires architectural remapping, not syntax substitution. CUDA encodes decomposition, locality, and synchronization through threads, blocks, and memory accesses; the Cerebras Software Language (CSL) requires explicit placement, distributed SRAM, fabric communication, event-driven tasks, and host/device contracts. We present FABRICA-Bench, 49 paired CUDA-to-CSL tasks, and FABRICA, an agentic framework combining target knowledge, execution, failure-directed repair, and correctness-gated optimization. On a fixed 28-task Level~1--3 core comparison with Claude Opus 4.8, FABRICA raises success from 6/28 to 26/28; 22 successful programs match or beat their CSL references. Across the 49-task coverage evaluation, 38 tasks produce a correct program; the final three tasks are evaluated over three seeds and pass 8/9 runs. For 27 generated/reference pairs with device-internal timing, geometric-mean speedup is 3.75$\times$ on the SDK simulator and 3.47$\times$ on WSE-3 hardware. With the executable workflow fixed, Claude Opus~4.8 passes 26/28 core tasks while the best open-weight model passes 2/28; retrieved Cerebras knowledge separately raises success from 1/15 to 7/15 on a Level~1--3 panel. These results identify base-model capability, target knowledge, execution feedback, and same-target measurement as central to cross-architecture kernel generation.

Problem

• CUDA
• CSL
• Architecture Porting
• Kernel Translation
• Optimization

Innovation

• CUDA-to-CSL translation
• Wafer-scale systems
• Agentic framework
• Execution feedback
• Failure-directed repair