A 33.6-136.2 TOPS/W Nonlinear Analog Computing-In-Memory Macro for Multi-bit LSTM Accelerator in 65 nm CMOS

TL;DR

This paper introduces a highly energy-efficient analog in-memory macro for LSTM accelerators, enabling direct nonlinear activation computation with high accuracy and robustness, significantly improving energy and area efficiency over existing solutions.

Contribution

It presents a novel nonlinear in-memory ADC and macro design for LSTM accelerators, achieving high energy efficiency and accuracy in analog computing-in-memory architectures.

Findings

Achieves 92.0% inference accuracy on keyword spotting

Demonstrates 2.2X higher energy efficiency than state-of-the-art

Shows robustness of ADC against temperature variations

Abstract

The energy efficiency of analog computing-in-memory (ACIM) accelerator for recurrent neural networks, particularly long short-term memory (LSTM) network, is limited by the high proportion of nonlinear (NL) operations typically executed digitally. To address this, we propose an LSTM accelerator incorporating an ACIM macro with reconfigurable (1-5 bit) nonlinear in-memory (NLIM) analog-to-digital converter (ADC) to compute NL activations directly in the analog domain using: 1) a dual 9T bitcell with decoupled read/write paths for signed inputs and ternary weight operations; 2) a read-word-line underdrive Cascode (RUDC) technique achieving 2.8X higher read-bitline dynamic range than single-transistor designs (1.4X better over conventional Cascode structure with 7X lower current variation); 3) a dual-supply 6T-SRAM array for efficient multi-bit weight operations and reducing both bitcell count (7.8X) and latency (4X) for 5-bit weight operations. We experimentally demonstrate 5-bit NLIM ADC for approximating NL activations in LSTM cells, achieving average error <1 LSB. Simulation confirms the robustness of NLIM ADC against temperature variations thanks to the replica bias strategy. Our design achieves 92.0% on-chip inference accuracy for a 12-class keyword-spotting task while demonstrating 2.2X higher system-level normalized energy efficiency and 1.6X better normalized area efficiency than state-of-the-art works. The results combine physical measurements of a macro unit-accounting for the majority of LSTM operations (99% linear and 80% nonlinear operations)-with simulations of the remaining components, including additional LSTM and fully connected layers.