LLMulator: Generalizable Cost Modeling for Dataflow Accelerators with Input-Adaptive Control Flow

2025-08-25

AI Summary

Existing methods struggle to universally model architecture-, application-, and input-dependent control flow, resulting in low accuracy and slow speed for dataflow accelerator performance prediction—hindering efficient design-space exploration. This paper proposes the first general-purpose performance prediction framework for dataflow accelerators: it models performance values as categorical token sequences, integrating large language models for program semantic encoding and program-structure awareness; introduces a reinforcement learning–driven dynamic calibration mechanism to adapt to input-variable control flow; and designs a multi-level dataflow pattern generation scheme with progressive data augmentation to enhance cross-architecture generalization. Experiments demonstrate that the framework achieves range-agnostic prediction and confidence estimation on unseen applications, reducing loop-cycle prediction error to 11.2%—a 9.7% improvement over static baselines.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Planning, Routing, and Scheduling: Planning with Language Models
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Accurate and fast performance prediction for dataflow-based accelerators is vital for efficient hardware design and design space exploration, yet existing methods struggle to generalize across architectures, applications, and input-dependent control flows. We present LLMulator, a progressive numeric modeling framework leveraging the program semantic knowledge of pre-trained large language models (LLMs) for robust, hardware- and application-aware prediction. Our numeric model treats performance values as categorical token sequences, enabling range-agnostic estimates and confidence-aware predictions for unseen applications. To handle input-dependent control flows, we introduce a reinforcement learning-based dynamic calibration method, reducing cycle prediction error by 9.7% over static models and converging to 11.2% error after a few iterations. For cross-hardware generalization, we develop a progressive data augmentation strategy that generates diverse datasets covering multi-level dataflow structures, memory parameters, and loop mapping primitives, significantly boosting prediction accuracy across architectures and configurations.

Problem

• Accurate performance prediction for dataflow accelerators across architectures
• Handling input-dependent control flows in hardware performance modeling
• Generalizing performance models across diverse applications and configurations

Innovation

• Leveraging LLMs for hardware-aware performance prediction
• Reinforcement learning for input-dependent control flow
• Progressive data augmentation for cross-hardware generalization