Quantization (signal)

MEGA.mini: An Energy-Efficient NPU Leveraging a Novel Big/Little Core With Hybrid Input Activation for Generative AI Acceleration

MEGA.mini: An Energy-Efficient NPU Leveraging a Novel Big/Little Core With Hybrid Input Activation for Generative AI Acceleration 150 150

Abstract:

This article presents a processor for the acceleration of generative AI (GenAI) based on a novel heterogeneous core architecture called MEGA.mini. The processor introduces three algorithmic features: 1) fixed-point (FXP) and floating-point (FP) hybrid input activation (IA) representation; 2) a delayed-statistics-based normalization (NORM); and 3) conditional polynomial-based nonlinear activation (NLA) approximation. These …

View on IEEE Xplore

A Calibration-Free Pipelined-SAR ADC With Cross-Stage Gain-Mismatch Error Shaping and Inherent Noise Shaping

A Calibration-Free Pipelined-SAR ADC With Cross-Stage Gain-Mismatch Error Shaping and Inherent Noise Shaping 150 150

Abstract:

This article presents a calibration-free pipelined-successive-approximation-register (SAR) analog-to-digital converter (ADC) based on the proposed cross-stage gain-mismatch-error shaping (CS-GMES) mechanism. The CS-GMES is realized by including the entire 2nd stage into MES operation to unify the gain error and the 2nd-stage mismatch error. A feedback capacitor provides cross-stage connection and mismatch …

View on IEEE Xplore

A 14-b Energy-Efficient BW/Power Scalable CTDSM With a Frequency-Controlled Current Source

A 14-b Energy-Efficient BW/Power Scalable CTDSM With a Frequency-Controlled Current Source 150 150

Abstract:

This work presents a 14-bit energy-efficient bandwidth (BW)/power scalable continuous-time delta–sigma modulator (CTDSM) for sensor interfaces in IoT applications. To ensure low noise for small input signals and achieve BW/power scalability, it is built around Gm-C integrators biased via a linear frequency-controlled current source (FCCS). The FCCS …

View on IEEE Xplore

DPIM: A 2T1C eDRAM Transformer-in-Memory Chip With Sparsity-Aware Quantization and Heterogeneous Dense–Sparse Core

DPIM: A 2T1C eDRAM Transformer-in-Memory Chip With Sparsity-Aware Quantization and Heterogeneous Dense–Sparse Core 150 150

Abstract:

Transformer models have revolutionized artificial intelligence (AI) applications across various domains, but their increasing complexity poses significant challenges in terms of computational and memory demands. While processing-in-memory (PIM) paradigms have been adopted to address these limitations, existing PIM-based transformer accelerators still face hurdles such as: 1) focusing solely on optimizing attention …

View on IEEE Xplore

A 6.2b-ENOB 2.5 GS/s Flash-and-VCO-Based Subranging ADC Using a Residue Shifting Technique

A 6.2b-ENOB 2.5 GS/s Flash-and-VCO-Based Subranging ADC Using a Residue Shifting Technique 150 150

Abstract:

This letter presents a 7-bit pipelined subranging ADC that integrates a 3-bit flash ADC with a ring VCO-based quantizer. A resistor-ladder-based residue shifter (RLRS) replaces traditional residue amplifiers, efficiently shifting the residue voltage into the most linear region of the $K_{textrm {VCO}}$ , thereby eliminating the need for post-linearity calibration. …

View on IEEE Xplore