Keywords: Reinforcement Learning, Memory Mapping, Device Placement, Evolutionary Algorithms
Abstract: For deep neural network accelerators, memory movement is both energetically expensive and can bound computation. Therefore, optimal mapping of tensors to memory hierarchies is critical to performance. The growing complexity of neural networks calls for automated memory mapping instead of manual heuristic approaches; yet the search space of neural network computational graphs have previously been prohibitively large. We introduce Evolutionary Graph Reinforcement Learning (EGRL), a method designed for large search spaces, that combines graph neural networks, reinforcement learning, and evolutionary search. A set of fast, stateless policies guide the evolutionary search to improve its sample-efficiency. We train and validate our approach directly on the Intel NNP-I chip for inference. EGRL outperforms policy-gradient, evolutionary search and dynamic programming baselines on BERT, ResNet-101 and ResNet-50. We additionally achieve 28-78% speed-up compared to the native NNP-I compiler on all three workloads.
One-sentence Summary: We combine evolutionary and gradient-based reinforcement learning to tackle the large search spaces needed to map tensors to memory, yielding up to 78% speedup on BERT and ResNet on a deep learning inference chip.
Supplementary Material: zip
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics