Efficient Precision-Scalable Hardware for Microscaling (MX) Processing in Robotics Learning

TL;DR

This paper introduces a novel hardware design supporting all MX data types with shared exponents, significantly reducing memory and increasing training throughput for robotics learning at the edge.

Contribution

It presents a precision-scalable arithmetic unit and shared exponent support that overcome limitations of prior MX processing hardware, enabling more efficient on-device learning.

Findings

51% lower memory footprint

4x higher training throughput

comparable energy efficiency

Abstract

Autonomous robots require efficient on-device learning to adapt to new environments without cloud dependency. For this edge training, Microscaling (MX) data types offer a promising solution by combining integer and floating-point representations with shared exponents, reducing energy consumption while maintaining accuracy. However, the state-of-the-art continuous learning processor, namely Dacapo, faces limitations with its MXINT-only support and inefficient vector-based grouping during backpropagation. In this paper, we present, to the best of our knowledge, the first work that addresses these limitations with two key innovations: (1) a precision-scalable arithmetic unit that supports all six MX data types by exploiting sub-word parallelism and unified integer and floating-point processing; and (2) support for square shared exponent groups to enable efficient weight handling during backpropagation, removing storage redundancy and quantization overhead. We evaluate our design against Dacapo under iso-peak-throughput on four robotics workloads in TSMC 16nm FinFET technology at 400MHz, reaching a 51% lower memory footprint, and 4x higher effective training throughput, while achieving comparable energy efficiency, enabling efficient robotics continual learning at the edge.