Picasso: An Area/Energy-Efficient End-to-End Diffusion Accelerator with Hyper-Precision Data Type

This work presents Picasso, an end-to-end diffusion accelerator designed for enhancing the efficiency of diffusion-based machine learning models used in applications such as image and video generation, and inpainting. Picasso introduces a novel hyper-precision 8 (HYP8) data type and a reconfigurable...

Full description

Saved in:
Bibliographic Details
Main Authors: Yoo, Sungyeob, Ko, Geonwoo, Ham, Seri, Kim, Seeyeon, Chen, Yi, Kim, Joo-Young
Format: Conference Proceeding
Language:English
Subjects:
CMOS technology
Distance measurement
Energy efficiency
Hardware
Machine learning
Pipelines
Streaming media
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This work presents Picasso, an end-to-end diffusion accelerator designed for enhancing the efficiency of diffusion-based machine learning models used in applications such as image and video generation, and inpainting. Picasso introduces a novel hyper-precision 8 (HYP8) data type and a reconfigurable architecture designed to significantly enhance hardware efficiency, providing an extended dynamic range without sacrificing accuracy. It also features a unified engine that streamlines the processing of all non-matrix operations and employs sub-block pipeline scheduling to reduce overall latency. Fabricated in 28nm CMOS technology, this accelerator achieves an energy efficiency of 4.96 TOPS/W and a peak performance of 9.83 TOPS. Compared to previous works, Picasso demonstrates speedups ranging from 8.4× to 26.8× while also improving energy and area efficiency by 1.1× to 2.8× and 3.6× to 30.5×, respectively.
ISSN:2573-2048
DOI:10.1109/HCS61935.2024.10664910