EPIM: Efficient Processing-In-Memory Accelerators based on Epitome (2311.07620v3)
Abstract: The utilization of large-scale neural networks on Processing-In-Memory (PIM) accelerators encounters challenges due to constrained on-chip memory capacity. To tackle this issue, current works explore model compression algorithms to reduce the size of Convolutional Neural Networks (CNNs). Most of these algorithms either aim to represent neural operators with reduced-size parameters (e.g., quantization) or search for the best combinations of neural operators (e.g., neural architecture search). Designing neural operators to align with PIM accelerators' specifications is an area that warrants further study. In this paper, we introduce the Epitome, a lightweight neural operator offering convolution-like functionality, to craft memory-efficient CNN operators for PIM accelerators (EPIM). On the software side, we evaluate epitomes' latency and energy on PIM accelerators and introduce a PIM-aware layer-wise design method to enhance their hardware efficiency. We apply epitome-aware quantization to further reduce the size of epitomes. On the hardware side, we modify the datapath of current PIM accelerators to accommodate epitomes and implement a feature map reuse technique to reduce computation cost. Experimental results reveal that our 3-bit quantized EPIM-ResNet50 attains 71.59% top-1 accuracy on ImageNet, reducing crossbar areas by 30.65 times. EPIM surpasses the state-of-the-art pruning methods on PIM.
- Prime: A novel processing-in-memory architecture for neural network computation in reram-based main memory. ACM SIGARCH Computer Architecture News 44, 3 (2016), 27–39.
- Pim-prune: Fine-grain dcnn pruning for crossbar-based process-in-memory architecture. In 2020 57th ACM/IEEE Design Automation Conference (DAC). IEEE, 1–6.
- Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition. Ieee, 248–255.
- HAWQ-V2: Hessian Aware trace-Weighted Quantization of Neural Networks. Advances in neural information processing systems (2020).
- Hawq: Hessian aware quantization of neural networks with mixed-precision. In Proceedings of the IEEE/CVF International Conference on Computer Vision. 293–302.
- Device-circuit-architecture co-exploration for computing-in-memory neural accelerators. IEEE Trans. Comput. 70, 4 (2020), 595–605.
- ISAAC: A convolutional neural network accelerator with in-situ analog arithmetic in crossbars. ACM SIGARCH Computer Architecture News 44, 3 (2016), 14–26.
- Gibbon: efficient co-exploration of NN model and processing-in-memory architecture. In 2022 Design, Automation & Test in Europe Conference & Exhibition (DATE). IEEE, 867–872.
- An energy-efficient quantized and regularized training framework for processing-in-memory accelerators. In 2020 25th Asia and South Pacific Design Automation Conference (ASP-DAC). IEEE, 325–330.
- A compute-in-memory chip based on resistive random-access memory. Nature 608, 7923 (2022), 504–512.
- Neural epitome search for architecture-agnostic network compression. arXiv preprint arXiv:1907.05642 (2019).
- A configurable multi-precision CNN computing framework based on single bit RRAM. In 2019 56th ACM/IEEE Design Automation Conference (DAC). IEEE, 1–6.
- MNSIM 2.0: A behavior-level modeling tool for memristor-based neuromorphic computing systems. In Proceedings of the 2020 on Great Lakes Symposium on VLSI. 83–88.