IRONMAN: GNN-assisted Design Space Exploration in High-Level Synthesis via Reinforcement Learning
IRONMAN: GNN-assisted Design Space Exploration in High-Level Synthesis via Reinforcement Learning
复制标题
DOI：
10.1145/3453688.3461495
复制
发表时间：
2021-02
期刊：
影响因子：
--
通讯作者：
Nan Wu;Yuan Xie;Cong Hao
中科院分区：
文献类型：
--
作者：
Nan Wu;Yuan Xie;Cong Hao
尽管高级综合（HLS）工具取得了巨大的成功，我们观察到几个未解决的挑战：1）在HLS的编程风格的高层次抽象隐藏优化的机会; 2）现有的HLS工具不提供灵活的权衡不同的目标和约束; 3）最终RTL设计的实际质量是很难预测的。为此，我们提出了一个端到端的框架，IRONMAN。主要目标是实现灵活和自动化的设计空间探索（DSE），其可以在用户指定的约束下提供优化的解决方案，或者在不同目标（例如，资源类型、区域和等待时间）。IronMan由三个组件组成：GPP（一个高度精确的基于图神经网络的性能预测器），RLMD（一个基于学习的DSE引擎，探索优化的资源分配策略）和CT（一个代码Transformer，通过从原始HLS C/C++中提取数据流图来帮助RLMD和GPP）。实验结果表明：1）GPP具有较高的预测精度，在资源占用和时间上分别比HLS工具减少了10.9倍和5.7倍的预测误差; 2）RLMD获得的最优解或Pareto解分别比遗传算法和模拟退火算法提高了12.7%和12.9%; 3）IronMan可以找到完美匹配各种DSP约束的优化解决方案，DSP数量比HLS工具少2.54倍，延迟缩短高达6倍。IronMan也比元启发式技术和HLS工具快400倍。
Despite the great success of High-Level Synthesis (HLS) tools, we observe several unresolved challenges: 1) the high-level abstraction of programming styles in HLS conceals optimization opportunities; 2) existing HLS tools do not provide flexible trade-offs among different objectives and constraints; 3) the actual quality of the resulting RTL designs is hard to predict. To this end, we propose an end-to-end framework, IRONMAN. The primary goal is to enable a flexible and automated design space exploration (DSE), which can provide either optimized solutions under user-specified constraints, or Pareto trade-offs among different objectives (e.g., resource types, area, and latency). IronMan consists of three components: GPP (a highly accurate graph-neural-network-based performance predictor), RLMD (a reinforcement-learning-based DSE engine that explores the optimized resource allocation strategy), and CT (a code transformer that assists RLMD and GPP by extracting data flow graphs from original HLS C/C++). Experimental results show that, 1) GPP achieves high prediction accuracy, reducing prediction errors of HLS tools by 10.9X in resource usage and 5.7X in timing; 2) RLMD obtains optimized or Pareto solutions outperforming genetic algorithm and simulated annealing by 12.7% and 12.9%, respectively; 3) IronMan can find optimized solutions perfectly matching various DSP constraints, with 2.54X fewer DSPs and up to 6X shorter latency than those of HLS tools. IronMan is also up to 400X faster than meta-heuristic techniques and HLS tools.