# Accurate Sampling with Noisy Forces from Approximate Computing

## TL;DR

This paper introduces an algorithmic approach within approximate computing to accurately sample noisy forces in atomistic simulations, compensating for low-precision errors to obtain exact expectation values.

## Contribution

It proposes a novel method that rigorously corrects numerical inaccuracies from low-precision hardware in atomistic simulations using a modified Langevin equation.

## Key findings

- Effective compensation for low-precision errors demonstrated
- Exact expectation values achieved despite noisy forces
- Applicable to high-performance, energy-efficient computing hardware

## Abstract

In scientific computing, the acceleration of atomistic computer simulations by means of custom hardware is finding ever growing application. A major limitation, however, is that the high efficiency in terms of performance and low power consumption entails the massive usage of low-precision computing units. Here, based on the approximate computing paradigm, we present an algorithmic method to rigorously compensate for numerical inaccuracies due to low-accuracy arithmetic operations, yet still obtaining exact expectation values using a properly modified Langevin-type equation.