All-in-Memory Stochastic Computing using ReRAM

João Paulo C. de Lima 1,2 , Mehran Shoushtari Moghadam 3 , Sercan Aygun 4 , Jeronimo Castrillon 1,2,5 , M. Hassan Najafi 3 , and Asif Ali Khan 1 · 2025

Read All-in-Memory Stochastic Computing using ReRAM by João Paulo C. de Lima 1,2 , Mehran Shoushtari Moghadam 3 , Sercan Aygun 4 , Jeronimo Castrillon 1,2,5 , M. Hassan Najafi 3 , and Asif Ali Khan 1 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-08340.txt.

Description

João Paulo de Lima and co-authors run stochastic computing entirely inside memory chips in a 2025 paper. Stochastic computing trades exact digits for random bit streams, where simple gates do arithmetic and errors wash out. The catch has been making the streams: classic generators eat most of the system's power and area.

Resistive memory cells can mint true random numbers from their own physics. The team builds the whole flow in place: stream generation, the operations themselves, and the trip back to binary, all without shuttling data to a processor. The memory's flakiness, usually a flaw, meets a method built to shrug off bit errors. Tests on image tasks show clear throughput and energy wins over chip-based and earlier memory-based designs, at a small average cost in image quality.

Catalog record

CREATORJoão Paulo C. de Lima 1,2 , Mehran Shoushtari Moghadam 3 , Sercan Aygun 4 , Jeronimo Castrillon 1,2,5 , M. Hassan Najafi 3 , and Asif Ali Khan 1
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE35 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-08340

Browse the full library · About the Tower of Babel