ISMM 2025
as the organisers put it
Welcome to the home page of the 2025 ACM SIGPLAN International Symposium on Memory Management (ISMM 2025)! 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’25 will held in-person as part of PLDI’25, sharing the venue and activities. ISMM 2025 Program Update 2025-05-31: We have the program online with ten papers and two distinguished keynote speakers. Join us for insightful present ...
who it's for
garbage collector implementers, runtime engineers, systems researchers, compiler engineers, persistent memory engineers, machine learning engineers, embedded systems engineers, security researchers
topics
crash consistency for persistent memory; transparent huge page debloating; proactive memory tiering for DNNs; fully randomized pointer encoding; immediate cycle reclamation; lifetime dispersion and generational GC; production OpenJDK GC practices; MMTk integration for CRuby; MMTk heap interface for Julia; generational GC for PIM/DPUs; benchmarking LLMs for memory bugs; domain-specific memory management for PIM
| title | author | length |
|---|---|---|
| Compiler-Assisted Crash Consistency for PMEM | Yun Joon Soh, Sihang Liu, Steven Swanson, Jishen Zhao | 18m |
SSAPP is a compiler extension that transforms LLVM IR into two executables — a crash-consistent main program and an autogenerated recovery entry that computes… | ||
| EMD: Fair and Efficient Dynamic Memory De-bloating of Transparent Huge Pages | Parth Gangar, Ashish Panwar, K. Gopinath | 22m |
EMD is a kernel-level system for dynamic de-bloating of transparent huge pages that breaks huge pages into base pages, deduplicates zero-filled bases, and… | ||
| TierTrain: Proactive Memory Tiering for CPU-Based DNN Training | Sathvik Swaminathan, Sandeep Kumar, Aravinda Prasad, Sreenivas Subramoney | 18m |
TierTrain implements a proactive queuing daemon that uses PyTorch hooks and a first-epoch profile to evict and prefetch saved tensors across HBM/DRAM… | ||
| Fully Randomized Pointers | Sai Dhawal Phaye, Gregory J. Duck, Roland H. C. Yap, Trevor E. Carlson | 17m |
Proposes Fully Randomized Pointers (FRP), a pointer encoding that replaces machine addresses with a 40‑bit object ID plus randomized offset packed into 64… | ||
| Arborescent Garbage Collection: A Dynamic Graph Approach to Immediate Cycle Collection | Frédéric Lahaie-Bertrand, Léonard Oest O'Leary, Olivier Melançon, Marc Feeley, Stefan Monnier | 23m |
Embeds a spanning forest embedded in reference graph and maintains it on each pointer update so the runtime can perform immediate synchronous cycle… | ||
| Keynote: Industry GC Insights from OpenJDK | Erik Österlund | 59m |
OpenJDK work shows production GCs take about 10 years to production because collectors must integrate with JVMTI, compilers, class loading, debuggers and… | ||
| SecureMind: A Framework for Benchmarking Large Language Models in Memory Bug Detection(…) | Huanting Wang, Dejice Jacob, David Kelly, Yehia Elkhatib, Jeremy Singer, Zheng Wang | 19m |
A Python framework that automates dataset retrieval, cutoff-date filtering to avoid data leakage, probe orchestration and metric calculation, SecureMind… | ||
| Lifetime Dispersion and Generational GC: An Intellectual Abstract | Stephen Dolan | 24m |
Argues that lifetime dispersion (not “most objects die young”) predicts when generational GC helps, quantifies it with the Gini coefficient derived from the… | ||
| Reworking Memory Management in CRuby: A Practitioner Report | Kunshan Wang, Stephen M. Blackburn, Peter Zhu, Matt Valentine-House | 19m |
Reports integrating MMTk into CRuby by refactoring a non‑modular GC into a pluggable interface, implementing four stages: no-GC, non‑moving, copying… | ||
| Reconsidering Garbage Collection in Julia: A Practitioner Report | Luis Eduardo de Souza Amorim, Yi Lin, Stephen M. Blackburn, Diogo Netto, Gabriel Baraldi, Nathan Daly, Antony L. Hosking, Kiran Pamnany, Oscar Smith | 19m |
A third‑party heap interface for Julia is implemented and integrated with MMTk, enabling multiple GC plans (no‑GC, IMIX/EMIX non‑moving variants and sticky… | ||
| Gray-in-Young: A Generational Garbage Collection for Processing-in-Memory | Ryu Morimoto, Kazuki Ichinose, Tomoharu Ugawa | 20m |
A generational GC for UPMEM DPUs keeps the young generation in SPM and copies objects to old DRAM only after pointer updates, using reserve-and-scan to batch… | ||
| Keynote: Domain-Specific Memory Management? | John Kim | 52m |
Argues that PIM must be treated as PIM-as-memory not an accelerator, because treating it as an accelerator wastes capacity (Samsung PIM prototypes remove ~1/3… | ||
| nothing matches | ||