ISMM 2024
as the organisers put it
Welcome to the home page of the 2024 ACM SIGPLAN International Symposium on Memory Management (ISMM 2024)! 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’24 will held in-person as part of PLDI’24, sharing the venue and activities. Update! We’re thrilled to announce two distinguished keynote speakers for ISMM 2024. Join us for their insightful presentations on the future of memor ...
who it's for
memory management researchers, runtime engineers, garbage collector implementers, machine learning infrastructure engineers, hardware architects, embedded systems engineers, systems security engineers
topics
SSD offload for embeddings; GPU Direct Storage prefetching; simulated annealing allocation ordering; PyTorch caching fragmentation reduction; function-summaries concolic testing; reference counting for frozen SCCs; batching slab-directed frees; message-passing allocator optimizations; one-pass heap compaction; processing-in-memory for AI (AiM); managed virtual memory for microcontrollers; ephemeron-based lifetime profiling
| title | author | length |
|---|---|---|
| ESPN: Memory-Efficient Multi-Vector Information Retrieval | Susav Shrestha, Narasimha Reddy, Zongwang Li | 21m |
ESPN offloads multi-vector reranking embeddings from RAM to SSDs and uses GPU Direct Storage plus a software prefetcher to overlap compute and IO, reducing… | ||
| A heuristic algorithm for periodic memory allocation with little fragmentation to train(…) | Akifumi Imanishi, Zijian Xu | 21m |
Claims an offline simulated-annealing heuristic that reorders allocation topological ordering to cut PyTorch caching fragmentation from 29.5% wasted by PyTorch… | ||
| SSRD: Shapes and Summaries for Race Detection in Concurrent Data Structures | Xiaofan Sun, Rajiv Gupta | 17m |
SSRD implements summarization-guided concolic testing that generates function summaries from unit testing (capturing pointer-pointee relations and symbolic… | ||
| Reference Counting Deeply Immutable Data Structures with Cycles | Matthew J. Parkinson, Sylvan Clebsch, Tobias Wrigstad | 19m |
Claim: deeply frozen object graphs are collectible by reference counting by collapsing strongly connected components and maintaining single reference counter… | ||
| BatchIt: Optimizing Message-Passing Allocators for Producer-Consumer Workloads | Nathaniel Wesley Filardo, Matthew J. Parkinson | 20m |
Adds a small per-slab associative cache of remote frees that batches objects heading to the same slab, implemented for mimalloc and snmalloc implementations… | ||
| The One Pass (OP) Compactor (an intellectual abstract) | Tomer Cory, Erez Petrank | 18m |
The One Pass Compactor claims full mark-compact heap compaction using a single heap pass by combining pointer threading with an auxiliary address mapping so… | ||
| Keynote: Enabling the AI Revolution: Next-Generation Memory Solutions for AI Computing(…) | Euicheol Lim | 58m |
Claims a working GDDR6-based processing‑in‑memory chip called AiM that puts a lightweight GEMM accelerator inside each bank, with 0.5 TB/s per chip and 1 GB… | ||
| Session: ISMM - ISMM: Keynote 2 - Dr. Euicheol Lim | Euicheol Lim | 02m |
Shows that mainstream LLMs drive demand for HBM and expose GPU bandwidth and latency limits using ChatGPT as an example, arguing current GPU-based stacks hit a… | ||
| A Managed Memory System for micro controllers with NOR flash memory | Akira Inoue, Tomoharu Ugawa, Shigeru Chiba | 15m |
Prototype implements a managed virtual memory for microcontrollers (mruby on Raspberry Pi Pico with 264 KB SRAM and 2 MB NOR flash) that combines an eager… | ||
| Supporting Trusted Virtual Machines with Hardware-based Secure Remote Memory | Taekyung Heo, Seunghyo Kang, Sanghyeon Lee, Soojin Hwang, Joongun Park, Jaehyuk Huh | 19m |
Proposes TDM, a hardware engine that pairs with trusted VM features to encrypt, authenticate, and remotely map pages at page-granularity, perform ownership… | ||
| Finalization-Based Object Lifetime Profiling | Sebastian Jordan Montaño, Guillermo Polito, Stephane Ducasse, Pablo Tesone | 15m |
FiLiP instruments every allocation, attaches an ephemeron to record finalization time, and supports sampling (100, 50, 1, 0.1%) and optional pretenuring… | ||
| Garbage Collection for Mostly Serialized Heaps | Chaitanya S. Koparkar, Vidush Singhal, Aditya Gupta, Mike Rainey, Michael Vollmer, Artem Pelenitsyn, Sam Tobin-Hochstadt, Milind Kulkarni, Ryan R. Newton | 21m |
A hybrid generational collector for mostly-serialized heaps implements a nursery of mini-regions with bump allocation and a copying young collector that… | ||
| Memory Management on Mobile Devices | Kunal Sareen, Stephen M. Blackburn, Sara S. Hamouda, Lokesh Gidra | 20m |
Evaluates Android Runtime garbage collection with a reproducible framework, curated 10 real Android apps benchmarked via UI Automator, and compares DaCapo plus… | ||
| Keynote: Memory of Past and Future – Wasm’s Evolving Model of Store | Andreas Rossberg | 01h 06m |
Wasm’s store evolves from a single flat byte memory to a multi-component store: ref types and generalized tables keep opaque pointers out of linear memory, GC… | ||
| Session: ISMM - Welcome from the chairs | Michael D. Bond, Jae W. Lee, Hannes Payer | 07m |
Conference logistics and program details include lunch at 12:20 and a welcome reception at 18:00; proceedings are already in the ACM Digital Library… | ||
| nothing matches | ||