Study of CXL Memory Sharing with FamFS and its Use cases

CXL 3.0 introduces disaggregated memory for pooling <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> and sharing capabilities. Memory sharing can provide major benefits, including data de-duplication for scale-out jobs, making larger problems fit in memory, and avoidance of sharding and shuffling and their related complexity and overhead. In this paper we discuss ways to deploy scale-out memory sharing and the challenges involved. We introduce the Fabric Attached Memory File System (Famfs [3]), an open source file system access method for scale-out shared memory across multiple systems. We also present and benchmark examples of a database stored in shared CXL memory and concurrently queried by a cluster of systems. We finally analyze the benchmark results and compare it with system-ram + ssd based performance.

Authors

Institutions

Publication Details

Published
2024-12-18
DOI
https://doi.org/10.1109/hipcw63042.2024.00022
Primary Topic
Advanced Computational Techniques and Applications
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Study of CXL Memory Sharing with FamFS and its Use cases

Ramesh Aravind, Groves John
Advanced Computational Techniques and Applications
article

Study of CXL Memory Sharing with FamFS and its Use cases

Ramesh Aravind, Groves John
article en

Abstract

CXL 3.0 introduces disaggregated memory for pooling <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> and sharing capabilities. Memory sharing can provide major benefits, including data de-duplication for scale-out jobs, making larger problems fit in memory, and avoidance of sharding and shuffling and their related complexity and overhead. In this paper we discuss ways to deploy scale-out memory sharing and the challenges involved. We introduce the Fabric Attached Memory File System (Famfs [3]), an open source file system access method for scale-out shared memory across multiple systems. We also present and benchmark examples of a database stored in shared CXL memory and concurrently queried by a cluster of systems. We finally analyze the benchmark results and compare it with system-ram + ssd based performance.

Micron (United States) (US), Micron Biomedical (United States) (US)
Openalex Percentile: Top 75%
Advanced Computational Techniques and Applications
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.