Logo image
A 22nm 96Mb 50.6-to-90.2TFLOPS/W Non-Linear MLC ReRAM CIM Macro with High-Retention for Mamba/Transformer/CNN
會議論文集

A 22nm 96Mb 50.6-to-90.2TFLOPS/W Non-Linear MLC ReRAM CIM Macro with High-Retention for Mamba/Transformer/CNN

Hung-Hsi Hsu, Win-San Khwa, Yao-Kai Yeh, Chih-Ling Wu, Chang-Yuan Chen, Yen-Che Huang, Yen-Hua Lin, Cheng-Feng Chang, Yen-Tung Shao, Jen-Chun Tien, …
Digest of technical papers - IEEE International Solid-State Circuits Conference
IEEE International Solid State Circuits Conference
01/01/2026
Web of Science ID: WOS:001760995900214

摘要

Computer Science Computer Science, Artificial Intelligence Computer Science, Hardware & Architecture Computer Science, Interdisciplinary Applications Engineering Engineering, Electrical & Electronic Science & Technology Technology
We present a DTCO-designed high-retention multi-mode ReRAM nvCIM macro supporting linear MLC (L-MLC), non-linear MLC (NL-MLC), and SLC modes for Mamba, Transformer, and CNN workloads, addressing key challenges in inference accuracy, energy, and readout robustness. The proposed macro incorporates: 1) a reconfigurable compute mode for accuracy/efficiency tradeoff, 2) ISE-HF encoding to enhance input sparsity and save energy, 3) an S2R-Db ADC that enables high-yield, low-power NL-MLC readout. The design achieves 50.6-to-90.2TFLOPS/W (BF16) with ImageNet accuracy loss that is 79.17% lower than LMLC after 10 years of retention.

相關連結

指標

1 檢視次數

詳細資料

Logo image