Welcome to the home page of the 2023 ACM SIGPLAN International Symposium on Memory Management (ISMM 2023)!

ISMM is the premier forum dedicated to research in memory management, covering the areas of memory performance, allocator design, garbage collection, architectural support for memory management, persistent memories, emerging memory technologies, and more.

ISMM’23 will held in-person as part of FCRC’23, sharing the venue and activities with ten other top computer science conferences.

Code of Conduct

ISMM follows the ACM Policy Against Harassment at ACM Activities. Please familiarize yourself with the policy and guide for reporting unacceptable behavior.

Plenary
You're viewing the program in a time zone which is different from your device's time zone change time zone

Sun 18 Jun

Displayed time zone: Eastern Time (US & Canada) change

07:30 - 09:00
BreakfastPLDI Catering at Royal
07:30
90m
Other
Breakfast
PLDI Catering

09:00 - 10:20
09:00
5m
Day opening
Welcome to ISMM
ISMM 2023
Steve Blackburn Google and Australian National University, Erez Petrank Technion
09:05
60m
Talk
Keynote: A Brave New World for Memory Management Research
ISMM 2023
10:05
15m
Break
BreakSocial
ISMM 2023

10:20 - 11:00
ISMM: Session 2 - Application ScalabilityISMM 2023 at Magnolia 22
Chair(s): Steve Blackburn Google and Australian National University

#ismm-1020-session2-magnolia22 Discord icon small YouTube icon small

10:20
20m
Talk
Scaling Up Performance of Managed Applications on NUMA Systems
ISMM 2023
Orion Papadakis The University of Manchester, Andreas Andronikakis The University of Manchester, Nikos Foutris The University of Manchester, Michail Papadimitriou OctoML, Athanasios Stratikopoulos The University of Manchester, Foivos S. Zakkak Red Hat, Inc., Polychronis Xekalakis Nvidia, Christos Kotselidis Pierer Innovation / The University of Manchester, Foivos S. Zakkak Red Hat, Inc.
DOI
10:40
20m
Talk
Analyzing and Improving the Scalability of In-Memory Indices for Managed Search Engines
ISMM 2023
Aditya Chilukuri Australian National University, Shoaib Akram Australian National University
DOI
11:00 - 11:20
11:00
20m
Coffee break
Break
PLDI Catering

11:20 - 12:30
ISMM: Session 3 - Intellectual AbstractsISMM 2023 at Magnolia 22
Chair(s): Michael D. Bond Ohio State University, USA

#ismm-1120-session3-magnolia22 Discord icon small YouTube icon small

11:20
20m
Talk
Memory Consistency Models for Program Transformations: An Intellectual Abstract
ISMM 2023
Akshay Gopalakrishnan McGill University, Clark Verbrugge McGill University, Canada, Mark Batty University of Kent, Clark Verbrugge McGill University, Canada
DOI
11:40
20m
Talk
Predicting Dynamic Properties of Heap Allocations using Neural Networks Trained on Static Code: An Intellectual Abstract
ISMM 2023
Christian Navasca UCLA, Martin Maas Google, Petros Maniatis Google, Hyeontaek Lim Google, Harry Xu University of California, Los Angeles (UCLA)
DOI
12:00
20m
Talk
The Unexpected Efficiency of Bin Packing Algorithms for Dynamic Storage Allocation in the Wild: An Intellectual Abstract
ISMM 2023
Christos Lamprakos National Technical University of Athens, Katholieke Universiteit Leuven, Sotirios Xydis National Technical University of Athens, Francky Catthoor IMEC, Katholieke Universiteit Leuven, Dimitrios Soudris National Technical University of Athens
DOI
12:20
10m
Awards
Best Paper Award
ISMM 2023

12:30 - 14:00
12:30
90m
Lunch
Lunch
PLDI Catering

14:00 - 15:20
ISMM: Session 4 - Allocations and Garbage CollectionISMM 2023 at Magnolia 22
Chair(s): Tony Hosking Australian National University

#ismm-1400-session4-magnolia22 Discord icon small YouTube icon small

14:00
20m
Talk
Concurrent GCs and Modern Java Workloads: A Cache PerspectiveBest Paper Award
ISMM 2023
Maria Carpen-Amarie Huawei Zurich Research Center, Switzerland, Georgios Vavouliotis Huawei Zurich Research Center, Switzerland, Konstantinos Tovletoglou Huawei Zurich Research Center, Switzerland, Boris Grot University of Edinburgh, UK, Rene Mueller Huawei Zurich Research Center, Switzerland
DOI
14:20
20m
Talk
Wait-Free Weak Reference Counting
ISMM 2023
Matthew J. Parkinson Azure Research, Microsoft, UK, Sylvan Clebsch Azure Research, Ben Simner
DOI
14:40
20m
Talk
NUMAlloc: A Faster NUMA Memory Allocator
ISMM 2023
Hanmei Yang University of Massachusetts Amherst, Xin Zhao University of Massachusetts Amherst, Jin Zhou University of Massachusetts Amherst, Wei Wang University of Texas at San Antonio, USA, Sandip Kundu University of Massachusetts Amherst, Bo Wu Colorado School of Mines, Hui Guan University of Massachusetts, Amherst, Tongping Liu University of Massachusetts at Amherst
DOI
15:00
20m
Talk
Picking a CHERI Allocator: Security and Performance Considerations
ISMM 2023
Jacob Bramley Arm, Dejice Jacob University of Glasgow, UK, Andrei Lascu King's College London, Jeremy Singer University of Glasgow, Laurence Tratt King's College London, Andrei Lascu King's College London
DOI Pre-print
15:30 - 16:00
15:30
30m
Coffee break
Break
PLDI Catering

16:00 - 17:20
ISMM: Session 5 - MiscellaneousISMM 2023 at Magnolia 22
Chair(s): Martin Maas Google

#ismm-1600-session5-magnolia22 Discord icon small YouTube icon small

16:00
20m
Talk
Blast from the Past: Least Expected Use (LEU) Cache Replacement with Statistical History
ISMM 2023
Sayak Chakraborti University of Rochester, Zhizhou (Chris) Zhang Uber Technologies, Noah Bertram Cornell University, Sandhya Dwarkadas University of Rochester, Chen Ding University of Rochester
DOI
16:20
20m
Talk
OMRGx: Programmable and Transparent Out-of-Core Graph Partitioning and Processing
ISMM 2023
Gurneet Kaur University of California, Riverside, Rajiv Gupta UC Riverside
DOI
16:40
20m
Talk
ZipKV: In-Memory Key-Value Store with Built-In Data Compression
ISMM 2023
Linsen Ma Rensselaer Polytechnic Institute, Rui Xie Rensselaer Polytechnic Institute, Tong Zhang Rensselaer Polytechnic Institute
DOI
17:00
20m
Talk
Flexible and Effective Object Tiering for Heterogeneous Memory Systems
ISMM 2023
Brandon Kammerdiener University of Tennessee, Jeffrey Zachariah McMichael University of Tennessee, Michael Jantz University of Tennessee, Kshitij Doshi Intel Corporation, Terry Jones Oak Ridge National Laboratory
DOI

Accepted Papers

Title
Analyzing and Improving the Scalability of In-Memory Indices for Managed Search Engines
ISMM 2023
DOI
Blast from the Past: Least Expected Use (LEU) Cache Replacement with Statistical History
ISMM 2023
DOI
Concurrent GCs and Modern Java Workloads: A Cache PerspectiveBest Paper Award
ISMM 2023
DOI
Flexible and Effective Object Tiering for Heterogeneous Memory Systems
ISMM 2023
DOI
Memory Consistency Models for Program Transformations: An Intellectual Abstract
ISMM 2023
DOI
NUMAlloc: A Faster NUMA Memory Allocator
ISMM 2023
DOI
OMRGx: Programmable and Transparent Out-of-Core Graph Partitioning and Processing
ISMM 2023
DOI
Picking a CHERI Allocator: Security and Performance Considerations
ISMM 2023
DOI Pre-print
Predicting Dynamic Properties of Heap Allocations using Neural Networks Trained on Static Code: An Intellectual Abstract
ISMM 2023
DOI
Scaling Up Performance of Managed Applications on NUMA Systems
ISMM 2023
DOI
The Unexpected Efficiency of Bin Packing Algorithms for Dynamic Storage Allocation in the Wild: An Intellectual Abstract
ISMM 2023
DOI
Wait-Free Weak Reference Counting
ISMM 2023
DOI
ZipKV: In-Memory Key-Value Store with Built-In Data Compression
ISMM 2023
DOI

Call for Papers

The 2023 ACM SIGPLAN International Symposium on Memory Management (ISMM 2023) is soliciting full-length submissions covering new work on all memory management related topics, as well as papers presenting confirmations or refutations of important prior results. In additional to regular papers, traditionally submitted to ISMM, we also invite submissions of the following kinds:

  • Surveys and comparative analyses that shed new light on previously published techniques.
  • Practitioner reports, describing experience with memory management in production. Such papers are not expected to provide novel research contributions, but they should not have been previously published.
  • Intellectual abstracts, where researchers share designs, algorithms, or theory that may be interesting to the memory management community, but not yet evaluated.

Please indicate whether the paper is a regular paper, a survey, a practitioner report, or an intellectual abstract, by using a subtitle. For example, for a regular paper, include on of the following on the line below the title line: \subtitle{This submission is a regular paper}, \subtitle{This submission is a survey}, \subtitle{This submission is a practitioner report}, or \subtitle{This submission is an intellectual abstract}.

ISMM 2023 will be colocated with PLDI 2023 at FCRC’23.

Areas of interest include but are not limited to:

  • Garbage collection algorithms and implementations
  • Memory allocation and de-allocation
  • Memory system design and analysis
  • Hardware support for memory management
  • Memory management for large-scale data-intensive systems
  • Novel memory architectures
  • Memory management at datacenter and cloud scales
  • Formal analysis and verification of memory management algorithms
  • Compiler analyses to aid memory management
  • Tools to analyze memory usage of programs
  • Empirical analysis of memory intensive programs
  • Formal analysis and verification of memory intensive programs
  • Memory management for machine learning systems
  • Programming and management of emerging or persistent memories

The symposium welcomes industry practitioners presenting their recent practice and findings in memory management related to real-world deployments.

Formatting Instructions

All papers must be submitted on-line in Portable Document Format (PDF).

Papers should be formatted according to the two-column ACM proceedings format. Each paper should have no more than 12 pages, excluding bibliography, in 10pt font. There is no limit on the page count for references. Each reference must list all authors of the paper (do not use et al). The citations should be in numeric style, e.g., [52]. Submissions should be in PDF format and printable on US Letter and A4 sized paper. These requirements are all the same as in the previous years.

Papers that exceed the length requirement or deviate from the expected format will be rejected.

Make sure that figures and tables are legible, even after the paper is printed in gray-scale.

Appendices should not be part of the paper, but should be submitted as supplementary material. Supplementary material should also be anonymized, as described below.

As explained in more detail at http://www.sigplan.org/Resources/Author, authors should use the sigplan subformat of the acmart format. Please note the following: The first two lines should be:

\documentclass[sigplan,10pt,review,anonymous]{acmart}
\settopmatter{printfolios=true,printccs=false,printacmref=false}

The default citation style is numeric. Do not mess with the class file or settings to try to sneak in additional space. (Conversely, you may toggle the printccs and printacmref flags if you wish, but these changes will consume space.) Do not use the PACMPL files or format; ISMM is not using them. However, the template files were designed to make migrating a paper from one format to the other as simple as possible.

Double-Blind Reviewing

ISMM uses double-blind reviewing. This means that author names and affiliations must be omitted from the submission. Additionally, if the submission refers to prior work done by the authors, that reference should be made in third person. These are firm submission requirements. Any supplementary material must also be anonymized.

PLDI’s FAQ on Double-Blind Reviewing clarifies the policy for the most common scenarios. But there are many gray areas and trade-offs. If you have any doubts about how to interpret the double blind rules, please contact the Program Chair. If in any doubt you should contact the Program Chair for complex cases that are not fully covered by the FAQ.

Declaring Conflicts

When submitting the paper, you will need to declare potential conflicts. Conflicts should be declared between an adviser and an advisee (e.g., Ph.D., post-doc). Other conflicts include institutional conflicts, financial conflicts of interest, friends or relatives, or any recent co-authors on papers and proposals (last 2 years).

Please do not declare spurious conflicts: such incorrect conflicts are especially harmful if the aim is to exclude potential reviewers, so spurious conflicts can be grounds for rejection. If you are unsure about a conflict, please consult the Program Chair.

Publication Date

AUTHORS TAKE NOTE: The official publication date is the date the proceedings are made available in the ACM Digital Library. This date may be up to two weeks prior to the first day of your conference. The official publication date affects the deadline for any patent filings related to published work. (For those rare conferences whose proceedings are published in the ACM Digital Library after the conference is over, the official publication date remains the first day of the conference.)

Author Response

The author response process will occur in April 2023 (see Important Dates on the main page), and will give the authors an opportunity to respond to factual errors in reviews before the Program Committee meets to make its decisions. The committee may, but need not, respond to the author response or revise reviews at or after the committee meeting.

Acknowledgements

This call-for-papers is an adaptation and evolution of content from previous instances of ISMM and PLDI. We are grateful to prior organizers for their work, which is reused here.

ISMM’23 Speaker’s Guide

This document is for those presenting a paper at ISMM’23.

Congratulations on having your paper accepted at ISMM’23! This document will help ensure your presentation runs smoothly and has the best possible audience impact. Please read it in its entirety.

Checklist

Before ISMM:

  • Prepare and practice.
  • Ensure your talk runs for no more than 16 minutes.
  • Sign up for Discord (details here).
  • Upload a backup copy of your talk slides (details below).
  • Check the program to establish when and where your talk will be.
  • Ensure you have a HDMI adaptor for your device.

Before your talk:

  • Familiarize yourself with the room you will be speaking in (Magnolia 22).
  • Find and introduce yourself to your session chair.
  • Find the Discord channel for your session, so you can monitor questions.
  • Attend the video check before your session (details below).
  • Ensure that you are in the room no later than 5 minutes before your session.

After your talk:

  • Expect questions from the floor and session chair.
  • Once you’ve completed your talk and Q&A, monitor Discord for follow-up questions.

Preparing Your Talk

Your work will have a greater impact if you’re well prepared.

It is very important that you run to schedule. The ISMM’23 schedule is extremely tight, with hard stops imposed by FCRC scheduling. Session chairs have been asked to stick rigidly to the schedule.

Guidelines

  1. Your talk should run for no more than 16 minutes , uninterrupted. This gives you about three minutes for questions and one minute for speaker change-over.
  2. Your talk should be prepared for the standard 16:9 widescreen ratio. If your talk is in a different ratio, at best it will be pillarboxed, wasting screen real estate and diminishing impact, and at worst, it won’t display correctly.
  3. You will present your talk from a lectern, using a fixed lectern mic.
  4. You will need to provide your talk ahead of time in either pdf or powerpoint.
  5. If you have an embedded video in your presentation, please inform the video team during your mandatory video check before your session.

Uploading Your Presentation

As an insurance against technical failures, we ask all speakers to make a backup copy of their presentation available to the video team by uploading it the day before the session. You’re welcome to upload fresh copies at any time.

  • Format: your presentation must be saved as a powerpoint or pdf file (sorry!)
  • Naming: you must use your paper ID as your file name (e.g. <ismm paper ID>.[ppt,pptx,pdf])
  • Location: please use this link: https://bit.ly/pldi23upload

This requirement gives you assurance that if some major technical problem were to arise (such as a failure of your laptop), you will still be able to give your talk. If you do not make your presentation available in advance, and significant technical problems arise, we may have to shorten your presentation to keep to our tight schedule.

The requirement for you to use pdf or powerpoint for your backup copy is a pragmatic tradeoff. These slides will only be used in case of a technical emergency. We want to have the highest possible assurance that they will work without fuss on a third party device should such an emergency occur. If you use Google slides, Keynote, or some other software, please use the export feature to create either powerpoint or pdf backups.

If you elect not to upload a backup copy, please understand that this limits our volunteers’ capacity to assist you if a technical problem arises when you give your presentation.

Advice

There are many excellent sources of advice on giving good talks, including from Simon Peyton Jones, Michael Hicks, Michael Ernst, Ranjit Jhala, and Derek Dreyer. Make good use of these!

Mandatory Video Check

All speakers are required to be in the room and check in with their session chair and the video team no later than 5 minutes before their session starts. Please note that due to our very tight schedule, speakers who fail to upload their talk in advance and/or fail to attend the video check 5 minutes before their session may have their talk cut short if technical issues arise.

Q&A

If you stick to the above schedule you will have about 3 minutes for questions. The in-room audience will be able to ask questions via a queue at a single microphone on a mic stand in the center of the room. In-room attendees and remote attendees will also be able to ask questions via Discord. Your session chair will monitor questions on Discord and might ask questions as they see them appearing there.

It is good practice, as the speaker, to repeat your understanding of the question before providing your answer. This is particularly important when time is tight because it reduces opportunities for time being wasted on account of a misunderstanding.

Once your talk is finished, please go on to Discord and respond to any questions or follow-up questions that appear there.

Remote Audience

Your talk will be streamed to Discord and YouTube. Your remote audience will be able to write questions in the Discord channel created for your session (they won’t be able to ask questions via audio or video). They should see your slides, and a video feed of you speaking. As mentioned above, your session chair may relay questions from Discord.

Remote Presenters

ISMM’23 authors are expected to be present at the conference.