# MASR: A Modular Accelerator for Sparse RNNs

arXiv: https://arxiv.org/abs/1908.08976 · HTML: https://mlsanity.com/paper/1908.08976/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Aug 2019.

## Built on by (most cited first, 8 of 8)

- [Google Neural Network Models for Edge Devices: Analyzing and Mitigating Machine Learning Inference Bottlenecks](https://mlsanity.com/paper/2109.14320/) (2021)
- [S2TA: Exploiting Structured Sparsity for Energy-Efficient Mobile CNN Acceleration](https://mlsanity.com/paper/2107.07983/) (2021)
- [Circa: Stochastic ReLUs for Private Deep Learning](https://mlsanity.com/paper/2106.08475/) (2021)
- [RecPipe: Co-designing Models and Hardware to Jointly Optimize Recommendation Quality and Performance](https://mlsanity.com/paper/2105.08820/) (2021)
- [Mitigating Edge Machine Learning Inference Bottlenecks: An Empirical Study on Accelerating Google Edge Models](https://mlsanity.com/paper/2103.00768/) (2021)
- [Exploiting Beam Search Confidence for Energy-Efficient Speech Recognition](https://mlsanity.com/paper/2101.09083/) (2021)
- [Structured in Space, Randomized in Time: Leveraging Dropout in RNNs for Efficient Training](https://mlsanity.com/paper/2106.12089/) (2021)
- [Mixture-of-Rookies: Saving DNN Computations by Predicting ReLU Outputs](https://mlsanity.com/paper/2202.04990/) (2022)

## Abstract (arXiv)

Recurrent neural networks (RNNs) are becoming the de facto solution for speech recognition. RNNs exploit long-term temporal relationships in data by applying repeated, learned transformations. Unlike fully-connected (FC) layers with single vector matrix operations, RNN layers consist of hundreds of such operations chained over time. This poses challenges unique to RNNs that are not found in convolutional neural networks (CNNs) or FC models, namely large dynamic activation. In this paper we present MASR, a principled and modular architecture that accelerates bidirectional RNNs for on-chip ASR. MASR is designed to exploit sparsity in both dynamic activations and static weights. The architecture is enhanced by a series of dynamic activation optimizations that enable compact storage, ensure no energy is wasted computing null operations, and maintain high MAC utilization for highly parallel accelerator designs. In comparison to current state-of-the-art sparse neural network accelerators (e.g., EIE), MASR provides 2x area 3x energy, and 1.6x performance benefits. The modular nature of MASR enables designs that efficiently scale from resource-constrained low-power IoT applications to large-scale, highly parallel datacenter deployments.