ICS 2024: International Conference on Supercomputing

• List of papers to be presented at ICS 2024
• ICS 2024's proceedings
• The proceedings will be open access during the conference and for up to one month after the conference. We hope you will take a look at the papers that interest you and engage in active discussions with the presenters.

June 4th (Tue), 2024

NOTICE

The workshops will be held in the International Science Innovation Building (Buildings numbered 69). The locations of the workshops and the registration desk are both DIFFERENT from those of the main conference. More details are available here.

	Room 1 Conference Rooms 5a/5b (5th floor)	Room 2 Meeting Rooms E/F (5th floor)
12:30 - 17:00	Registration (The registration desk is at Conference Room 1 on the 1st floor of the International Science Innovation Building. Please present your photo ID (passport, etc.) when you check in.)
13:00 - 15:00	Workshop - 1 QCCC: International Workshop on Quantum Classical Cooperative Computing	Workshop - 2 2nd Workshop on FPGA Technologies for Adaptive Computing (FTAC 2024)
15:00 - 15:30	Coffee break
15:30 - 17:30	Workshop - 1 QCCC: International Workshop on Quantum Classical Cooperative Computing	Workshop - 2 2nd Workshop on FPGA Technologies for Adaptive Computing (FTAC 2024)

June 5th (Wed), 2024

NOTICE

From June 5th, the registration desk will be located at entrance of the International Conference Hall, 2nd floor, Clock Tower Centennial Hall (Buildings numbered 3).

ICS 2024 Day 1:

	Room A International Conference Hall (2nd floor)	Room B International Conference Hall (2nd floor)
12:00 - 18:30	Registration (The registration desk is at entrance of the International Conference Hall, 2nd floor, Clock Tower Centennial Hall. Please present your photo ID (passport, etc.) when you check in.)
13:00 - 13:20	Opening Remarks General Co-chairs: Valentina Salapura (Google), Prof. Kenji Kise (Tokyo Institute of Technology) Steering Committee Chair: Prof. Alex Veidenbaum (University of California Irvine) Program Co-chairs: Prof. Murali Annavaram (University of Southern California), Prof. Ana Lucia Varbanescu (University of Twente)
13:20 - 14:20	Session 1: Invited Talk Chair: Hironori Nakajo (Tokyo University of Agriculture and Technology) Prof. Mateo Valero (Director at Barcelona Supercomputing Center - Centro Nacional de Supercomputación) Title: European Supercomputers: Buy versus Build
14:20 - 14:40	Coffee break
14:40 - 16:20 (Each talk: max. 15 min presentation / 5 min Q&A)	Session 2: Best Paper Nominees Chair: Murali Annavaram (University of Southern California), Ana Lucia Varbanescu (University of Twente) DAWN: Matrix Operation-Optimized Algorithm for Shortest Paths Problem on Unweighted Graphs Arkade: k-Nearest Neighbor Search With Non-Euclidean Distances using GPU Ray Tracing (BEST PAPER AWARD) Shared Virtual Memory: Its Design and Performance Implications for Diverse Applications (BEST PAPER AWARD RUNNER-UP) FuseIM: Fusing Probabilistic Traversals for Influence Maximization on Exascale Systems An Autonomous Parallelization of Transformer Model Inference on Heterogeneous Edge Devices
16:20 - 16:40	Coffee break
16:40 - 18:00 (Each talk: max. 15 min presentation / 5 min Q&A)	Session 3A: Memory and Storage Systems Chair: Y.C. Tay (National University of Singapore) LCM: LLM-focused Hybrid SPM-cache Architecture with Cache Management for Multi-Core AI Accelerators HMComp: Extending Near-Memory Capacity using Compression in Hybrid Memory NUCAlloc: Fine-Grained Block Placement in Hashed Last-Level NUCA Caches Exploiting Vector Code Semantics for Efficient Data Cache Prefetching	Session 3B: Emerging supercomputing applications Chair: Mehmet Belviranli (Colorado School of Mines) Real-time High-resolution X-Ray Computed Tomography RayJoin: Fast and Precise Spatial Join Differentiating Set Intersections in Maximal Clique Enumeration by Function and Subproblem Size Optimizing Attention by Exploiting Data Reuse on ARM Multi-core CPUs
18:00 - 18:30	Break
18:30 - 20:20	Reception The reception will be held on the evening of June 5th to welcome all attendees with drinks and light refreshments.

June 6th (Thu), 2024

ICS 2024 Day 2:

	Room A International Conference Hall (2nd floor)	Room B International Conference Hall (2nd floor)
08:15 - 13:00	Excursions / Lunch The four-hour morning excursion on June 6th will take you to Kiyomizu-dera Temple and Nanzen-ji Temple in several chartered vehicles. Please be at the bus pick-up area on time. It is easy to find your way from Exit 6 of Gion-Shijo Station. After visiting these temples, we will have lunch at Okazaki Saryo Mameda near Heian Shrine. It is important to note that the number of lunches provided is limited. We, the organizing committee, estimate that we will be able to serve lunch to all participants, but depending on the number of participants, it may not be possible to serve lunch to everyone. In this case, we will serve lunch in the order in which you complete the online registration. We are sorry and hope you will understand.
12:50 - 17:30	Registration (The registration desk is at entrance of the International Conference Hall, 2nd floor, Clock Tower Centennial Hall. Please present your photo ID (passport, etc.) when you check in.)
13:00 - 14:00	Session 4: Keynote Talk Chair: Kenji Kise (Tokyo Institute of Technology) Naoki Shinjo (SVP, Head of Advanced Technology Development Unit Fujitsu Research FUJITSU LIMITED) Title: From Past to Future: The Legacy and Hypothesis of Supercomputing [pdf]
14:00 - 14:20	Coffee break
14:20 - 15:40 (Each talk: max. 15 min presentation / 5 min Q&A)	Session 5A: Reliability, dependability and availability Chair: Marco Minutoli (Pacific Northwest National Laboratory) Minimizing Coherence Error via Dynamic Decoupling Soft Error Resilience at Near-Zero Cost Understanding GPU Memory Corruption at Extreme Scale: The Summit Case Study Input Range Generation for Compiler-Induced Numerical Inconsistencies	Session 5B: Heterogeneous software: GPUs and domain specific accelerators Chair: Jun Yang (University of Pittsburgh) Accurate Computation of the Logarithm of Modified Bessel Functions on GPUs RDMA-Based Algorithms for Sparse Matrix Multiplication on GPUs Distributed Ranges: A Model for Distributed Data Structures, Algorithms, and Views Stencil Computation with Vector Outer Product
15:40 - 16:00	Coffee break
16:00 - 17:20 (Each talk: max. 15 min presentation / 5 min Q&A)	Session 6A: Cloud and ML Systems Efficiency Chair: Ryohei Kobayashi (University of Tsukuba) YMIR: A Scheduler for Foundation Model Fine-tuning Workloads in Datacenters DeepHYDRA: Resource-Efficient Time-Series Anomaly Detection in Dynamically-Configured Systems An Efficient and Scalable Approach to Build Co-occurrence Matrix for DNN's Embedding Layer Scheduling for Cyber-Physical Systems with Heterogeneous Processing Units under Real-World Constraints	Session 6B: Accelerator Designs Chair: Saday Sadayappan (University of Utah) SLIDEX: A Novel Architecture for Sliding Window Processing Quasar-ViT: Hardware-Oriented Quantization-Aware Architecture Search for Vision Transformers CLAY: CXL-based Scalable NDP Architecture Accelerating Embedding Layers NeOCNN: NTT-enabled Optical Convolution Neural Network Accelerator
18:30 - 21:00	Banquet / Dinners at Heian-Jingu The banquet will be held on the evening of June 6th at the HEIAN JINGU SHRINE Restaurant & Banquet Hall, which will offer traditional Japanese course meals that incorporate classic Japanese cooking techniques and feature local ingredients from Kyoto. This banquet has no specific dress code, so you don't need to dress up in formal attire. Feel free to attend and enjoy Kyoto's culinary delights and networking. It is important to note that the number of seats for the banquet is limited. We, the organizing committee, estimate that you will all be able to attend the banquet, but depending on the number of attendees, it may not be possible. In this case, we will provide the seat in the order in which you complete the online registration. We are sorry and hope you will understand.

June 7th (Fri), 2024

ICS 2024 Day 3:

	Room A International Conference Hall (2nd floor)	Room B International Conference Hall (2nd floor)
09:15 - 13:30	Registration (The registration desk is at entrance of the International Conference Hall, 2nd floor, Clock Tower Centennial Hall. Please present your photo ID (passport, etc.) when you check in.)
09:30 - 10:30	Session 7: Invited Talk Chair: Kiyofumi Tanaka (JAIST) Dr. Michael Gschwind (Director / Principal Engineer for PyTorch at Meta Platforms) Title: LLMs Everywhere: Acceleration from Servers to Mobile Devices in the Age of Generative AI
10:30 - 10:50	Coffee break
10:50 - 12:10 (Each talk: max. 15 min presentation / 5 min Q&A)	Session 8A: Supercomputing Software and Security Chair: Jens Domke (RIKEN R-CCS) sys-sage: A Unified Representation of Dynamic Topologies & Attributes on HPC Systems RTT: Reuse Time Tracking for Use-After-Free Detection Tile Size and Loop Order Selection using Machine Learning for Multi-/Many-Core Architectures Matrix-free SBP-SAT finite difference methods and the multigrid preconditioner on GPUs	Session 8B: Interconnects and Networks Chair: Satoshi Ohshima (Kyushu University) SmartFuse: Reconfigurable Smart Switches to Accelerate Fused Collectives in HPC Applications CommBench: Micro-Benchmarking Hierarchical Networks with Multi-GPU, Multi-NIC Nodes gZCCL: Compression-Accelerated Collective Communication Framework for GPU Clusters Enhanced UGAL Routing Schemes for Dragonfly Networks
12:10 - 13:30	Lunch (A lunch box is provided. You can choose between grilled meat, vegan or eel.)
13:30 - 14:50 (Each talk: max. 15 min presentation / 5 min Q&A)	Session 9A: Machine learning systems Chair: Youtao Zhang (University of Pittsburgh) A Coordinated Strategy for GNN Combining Computational Graph and Operator Optimizations AUTOSCHED: An Adaptive Self-configured Framework for Scheduling Deep Learning Training Workloads Sylva: Sparse Embedded Adapters via Hierarchical Approximate Second-Order Information Fasor: A Fast Tensor Program Optimization Framework for Efficient DNN Deployment	Session 9B: Software Design for Accelerators Chair: Miquel Pericàs (Chalmers University of Technology) SNOOPIE: A Multi-GPU Communication Profiler and Visualizer RadiK: Scalable and Optimized GPU-Parallel Radix Top-K Selection Accelerated Auto-Tuning of GPU Kernels for Tensor Computations FASTEN: Fast GPU-accelerated Segmented Matrix Multiplication for Heterogeneous Graph Neural Networks