A 40nm RRAM Compute-in-Memory Macro Featuring On-Chip Write-Verify and Offset-Cancelling ADC References
Authors:
- 13 September 2021
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 79-82
Abstract
Resistive random access memory (RRAM) based compute-in-memory (CIM) has shown great potentials for deep neural network (DNN) inference. Prior works generally used off-chip write-verify scheme to tighten the RRAM resistance distribution and used off-chip analog-to-digital converter (ADC) references to fine-tune partial sum quantization edges. Though off-chip techniques are viable for testing purposes, they are unsuitable for practical applications. This work presents an RRAM-CIM macro that features 1) on-chip write-verify to speed up initial weight programming and periodically refresh cells to compensate for resistance drift under stress, and 2) on-chip ADC reference generation that provides column-wise tunability to mitigate offsets induced by process variation to guarantee CIFAR-10 accuracy of >90%. The design is taped-out in TSMC N40 RRAM process, and achieves 36.4TOPS/W for $1\times 1\mathrm{b}$ MAC operations on VGG-8 network.
Keywords
Funding Information
- NSF (CCF-1903951)
- SRC
- DARPA
This publication has 6 references indexed in Scilit:
- Secure-RRAM: A 40nm 16kb Compute-in-Memory Macro with Reconfigurability, Sparsity Control, and Embedded Security2021
- 16.1 A 22nm 4Mb 8b-Precision ReRAM Computing-in-Memory Macro with 11.91 to 195.7TOPS/W for Tiny AI Edge Devices2021
- High-Density 3D Monolithically Integrated Multiple 1T1R Multi-Level-Cell for Neural Networks2020
- A Variation Robust Inference Engine Based on STT-MRAM with Parallel Read-Out2020
- 2-Bit-Per-Cell RRAM-Based In-Memory Computing for Area-/Energy-Efficient Deep LearningIEEE Solid-State Circuits Letters, 2020
- An N40 256K×44 embedded RRAM macro with SL-precharge SA and low-voltage current limiter to improve read and write performance2018