Logo image
A 8-b-Precision 6T SRAM Computing-in-Memory Macro Using Segmented-Bitline Charge-Sharing Scheme for AI Edge Chips
期刊文章   同儕審查

A 8-b-Precision 6T SRAM Computing-in-Memory Macro Using Segmented-Bitline Charge-Sharing Scheme for AI Edge Chips

Jian-Wei Su, Yen-Chi Chou, Ruhui Liu, Ta-Wei Liu, Pei-Jung Lu, Ping-Chun Wu, Yen-Lin Chung, Li-Yang Hong, Jin-Sheng Ren, Tianlong Pan, …
IEEE Journal of Solid-State Circuits, 卷.58(3), 頁碼.877-892
2023

摘要

Artificial intelligence;Artificial intelligence (AI);Capacitance;charge sharing;Common Information Model (computing);Common Information Model (electricity);computing-in-memory (CIM);inference;Logic gates;SRAM cells;static random access memory (SRAM);Transistors Electrical and Electronic Engineering

Advances in static random access memory (SRAM)-CIM devices are meant to increase capacity while improving energy efficiency (EF) and reducing computing latency ( $T_{mathrm{AC}})$ . This work presents a novel SRAM-CIM structure using: 1) a segmented-bitline charge-sharing (SBCS) scheme for multiply-and-accumulate (MAC) operations with low energy consumption and a consistently high signal margin across MAC values; 2) a bitline-combining (BL-CMB) scheme to reduce the number of analog-to-digital converters (ADCs) and, thereby, provide options in determining a tradeoff between EF and inference accuracy; 3) a source-injection local-multiplication cell (SILMC) connected to two types of global-bitline-switch to support the SBCS and BL-CMB schemes with consistent signal margin against process variation in transistors; and 4) prioritized-hybrid ADC to suppress area and power overhead for analog readout operations. We fabricated a 28-nm 384-kb SRAM-CIM macro using foundry-provided compact-6T cells supporting MAC operations with 16 accumulations of 8-b input and 8-b weight with near-full precision output (20 b). This macro achieved $T_{mathrm{AC}}$ of 7.2 ns and EF of 22.75 TOPS/W performing 8-b-MAC operations.

相關連結

指標

1 檢視次數

詳細資料

Logo image