skip to main content
10.1145/3698783acmconferencesBook PagePublication PagessospConference Proceedingsconference-collections
DIMES '24: Proceedings of the 2nd Workshop on Disruptive Memory Systems
ACM2024 Proceeding
Publisher:
  • Association for Computing Machinery
  • New York
  • NY
  • United States
Conference:
SOSP '24: ACM SIGOPS 30th Symposium on Operating Systems Principles Austin TX USA 3 November 2024
ISBN:
979-8-4007-1303-3
Published:
03 November 2024
Sponsors:
Recommend ACM DL
ALREADY A SUBSCRIBER?SIGN IN
Next Conference
October 13 - 16, 2025
Seoul , Republic of Korea
Reflects downloads up to 25 Dec 2024Bibliometrics
Skip Abstract Section
Abstract

About the Workshop on Disruptive Memory Systems (DIMES) Novel memory types are shattering our decades-old assumptions about the interface between hard- and software. Instead of volatile, passive and largely homogeneous main memory, current systems support non-volatile main memory, hybrid architectures combine classic memory modules with high-bandwidth memory, and disaggregated memory is no longer connected to the processor via the memory bus. In-memory and near-memory computing abandon the traditional von Neumann architecture and enable large numbers of parallel operations with the enormous potential for gains in performance and energy efficiency. The Workshop on Disruptive Memory Systems (DIMES) is a platform to discuss new visions, abstractions, and interfaces that will allow system software to exploit the opportunities of novel memory technologies in existing and future software. Our discussion focuses on system software of all types in various different computing domains (e.g., embedded, edge, cloud, and HPC).

Skip Table Of Content Section
SESSION: Let the Numbers Do the Talking - Benchmarks and Models
research-article
Performance Models for Task-based Scheduling with Disruptive Memory Technologies

Disruptive memory technologies break out of the memory pyramid and mandate specialized performance models and algorithms for optimal use. While the literature offers various - occasionally conflicting - models for individual technologies, few of these ...

research-article
Open Access
Demystifying Intel Data Streaming Accelerator for In-Memory Data Processing

In-memory data processing is the state-of-the-art approach for large-scale data analytics. For a more efficient processing and to be able to store more data in main memory, new hardware developments for memory such as Compute Express Link (CXL) and High-...

SESSION: There Must Be Order - Memory and Caches
research-article
To Keep or Not to Keep - The Volatility of Replacement Policy Metadata in Hybrid Caches

The field of Non-Volatile Memory (NVM) is one of the most actively researched topics regarding disruptive memory technologies. A promising application of NVM lies in the hybridization of caches, combining both NVM and conventional SRAM to realize trade-...

research-article
Open Access
Moses: Heap Partitioning for Semantic Data Tiering

The adoption of emerging memory technologies has long been impeded by the lack of interfaces communicating memory placement semantics not only between applications, system software, and diagnostics tools but also between multiple tenants in a system. ...

research-article
Open Access
The New Costs of Physical Memory Fragmentation

External fragmentation is becoming a serious problem again after paging temporarily solved it with its one-size-fits-all 4 KiB approach. The increasing adoption of mixed base, huge, and giant page sizes, DRAM energy-saving techniques, and memory ...

SESSION: A Silver Lining? - CXL Systems
research-article
Lupin: Tolerating Partial Failures in a CXL Pod

A compute express link (CXL) pod is a collection of hosts attached to a CXL memory module. It provides an opportunity to port single-host shared-memory programs to execute on multiple hosts in a CXL pod, where the ported application achieves higher ...

research-article
Open Access
Fundamental OS Design Considerations for CXL-based Hybrid SSDs

The first commercial implementations of CXL-based hybrid SSDs (i.e., SSDs that are both byte- and block-addressable) are looming on the horizon. Although previous works have conducted design studies on hardware concepts as well as potential use cases, ...

DEMONSTRATION SESSION: Cards on the Table - Demos
demonstration
Novel Memory Technologies for Multi-Tenant Exploratory Programming

The current deep learning trend has given new popularity to the Exploratory Programming style, which often comes with high and exclusive-use demands on compute hardware and memory resources. Yet, the underlying system software is unaware of the memory ...

Index terms have been assigned to the content through auto-classification.

Recommendations

Acceptance Rates

Overall Acceptance Rate 8 of 17 submissions, 47%
YearSubmittedAcceptedRate
DIMES '2317847%
Overall17847%