Heterogeneous 3-D Integration of Multitier Compute-in-Memory Accelerators: An Electrical-Thermal Co-Design
- 23 September 2021
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Electron Devices
- Vol. 68 (11) , 5598-5605
Abstract
Emerging nonvolatile memory (eNVM)-based compute-in-memory (CIM) accelerators have been proven in silicon for machine learning at the macrolevel. To fully unleash the system-level benefits, the heterogeneous 3-D integration (H3D) using through-silicon via (TSV) is a promising approach, to: 1) address the challenges of area-hungry peripheries in CIM accelerators; 2) solve the 2-D scaling challenges of eNVM; and 3) stack enormous amount of embedded memories that are required in state-of-the-art deep neural network models. This article presents an electrical-thermal co-design of multitier CIM accelerators, based on SRAM and/or eNVM, with hybrid technology nodes for logic and memory tiers. We benchmark the CIM accelerators on 8-bit ResNet-34 for ImageNet recognition, with layer-by-layer and pipelined schemes, respectively. By sweeping TSV diameter from $30~\mu \text{m}$ to 100 nm, we investigate the tradeoffs of system performance metrics (TOPS/W, TOPS, and TOPS/mm 2 ) and H3D challenges (thermal and IR-drop in power delivery). Finally, we find the sweet spot of TSV diameter for multitier H3D system is 1– $3~\mu \text{m}$ , to guarantee balanced area-overhead, performance, and IR-drop in power delivery. The extended benchmark framework is released on GitHub ( https://github.com/neurosim ) as an open-source tool for the research community.Keywords
Funding Information
- ASCENT, one of the SRC/DARPA JUMP Centers
- NSF (NSF-CCF-1903951)
This publication has 14 references indexed in Scilit:
- 2-Bit-Per-Cell RRAM-Based In-Memory Computing for Area-/Energy-Efficient Deep LearningIEEE Solid-State Circuits Letters, 2020
- Extreme Wafer Thinning and nano-TSV processing for 3D Heterogeneous Integration2020
- Training high-performance and large-scale deep neural networks with full 8-bit integersNeural Networks, 2020
- Optimizing Weight Mapping and Data Flow for Convolutional Neural Networks on Processing-in-Memory ArchitecturesIEEE Transactions on Circuits and Systems I: Regular Papers, 2019
- Impact of Non-Ideal Characteristics of Resistive Synaptic Devices on Implementing Convolutional Neural NetworksIEEE Journal on Emerging and Selected Topics in Circuits and Systems, 2019
- Through-Silicon Vias: Drivers, Performance, and InnovationsIEEE Transactions on Components, Packaging and Manufacturing Technology, 2016
- Thermal Design and Constraints for Heterogeneous Integrated Chip Stacks and Isolation Technology Using Air Gap and Thermal BridgeIEEE Transactions on Components, Packaging and Manufacturing Technology, 2014
- Novel through-silicon via technologies for 3D system integration2013
- Stress-driven 3D-IC placement with TSV keep-out zone and regularity study2010
- Analysis of the Induced Stresses in Silicon During Thermcompression Cu-Cu Bonding of Cu-Through-Vias in 3D-SIC Architecture2007