IEEE Cluster 2022 Program

Conference - Tuesday, September 6

8:30-9:15

Registration

9:15 - 10:30

HS6
HS7
HS8
HS9
Tutorial: Introduction to Research Data Management (RDM) with Hands-On for HPC Use Cases

10:30 - 11:00

Coffee Break

11:00 - 12:30

HS6
HS7
HS8
HS9
Tutorial: Introduction to Research Data Management (RDM) with Hands-On for HPC Use Cases

12:30 - 14:00

Lunch

14:00 - 15:30

HS6
HS7
HS8
HS9
Tutorial: Heterogeneous Programming in Modern C++ with SYCL

15:30 - 16:00

Coffee Break

16:00 - 17:30

HS6
HS7
HS8
HS9
Tutorial: Heterogeneous Programming in Modern C++ with SYCL

19:00

Networking Dinner — Kulturbrauerei Heidelberg
Registration required - limited seats. (Register here)

Conference - Wednesday, September 7

8:15-9:00

Registration

9:00 - 9:30

Cluster 2022 Opening
HS13

9:30 - 10:30

Keynote: Luca Benini, ETH Zurich
HS13
Mempools: The Rise of Tightly Coupled Processor Clusters
Chair: Trilce Estrada

10:30 - 11:00

Coffee Break

11:00 - 12:30 -- Parallel Sessions

Networking & Security
HS13
Chair: Alexandru Calotoiu
Bring the BitCODE - Moving Compute and Data in Distributed Heterogeneous Systems
Exploring Light-weight Cryptography for Efficient and Secure Lossy Data Compression
SKV: A SmartNIC-Offloaded Distributed Key-value Store

Scheduling and Multi-Tenancy
HS10
Chair: Dirk Pleiter
What does Inter-Cluster Job Submission and Execution Behavior Reveal to Us?
Matching-based Scheduling of Asynchronous Data Processing Workflows on the Computing Continuum
MRSch: Multi-Resource Scheduling for HPC

12:30 - 14:00

Poster Session and Lunch

14:00 - 15:30 -- Parallel Sessions

MPI
HS13
Chair: Sascha Hunold
A framework for hierarchical single-copy MPI collectives on multicore nodes
Deadlock Detection of MPI Program Based on Refined Match-sets

Runtimes
HS10
Chair: Bernd Mohr
Pythia: an oracle to guide runtime system decisions
Pushing the Boundaries of Small Tasks: Scalable Low-Overhead Data-Flow Programming in TTG
Distributed Continuation Stealing is More Scalable than You Might Think

15:30 - 16:00

Coffee Break

16:00 - 17:30 -- Parallel Sessions

MPI Collectives
HS13
Chair: Sascha Hunold
Fast(er) Construction of Round-optimal n-Block Broadcast Schedules
Lossy all-to-all exchange for accelerating parallel 3-D FFTs on hybrid architectures with GPUs
ACCLAiM: Advancing the Practicality of MPI Collective Communication Autotuning Using Machine Learning

Serverless & Virtual Networks
HS10
Chair: Jay Lofstead
Call Scheduling to Reduce Response Time of a FaaS System, Paweł Żuk
FaaSt: Optimize Makespan of Serverless Workflows in Federated Commercial FaaS
Last-mile Matters: Mitigating the Tail Latency of Virtualized Networks with Multipath Data Plane

17:30 - 19:30

Welcome Reception
Reception

Conference - Thursday, September 8

8:45-9:15

Registration

9:15 - 9:30

Announcements
HS13

9:30 - 10:30

Keynote: Kristal Michielsen, Jülich Supercomputing Centre
HS13
Integrating Quantum Computers in HPC Infrastructures
Chair: Felix Wolf

10:30 - 11:00

Coffee Break

11:00 - 12:30 -- Parallel Sessions

Applications
HS13
Chair: Tom Deakin
Towards Virtual Certification of Gas Turbine Engines With Performance-Portable Simulations
Hybrid Analysis of Fusion Data for Online Understanding of Complex Science on Extreme Scale Computers
High Performance Adaptive Physics Refinement to Enable Large-Scale Tracking of Cancer Cell Trajectory

I/O
HS10
Chair: Jay Lofstead
Be SMART, Save I/O: A Probabilistic Approach to Avoid Uncorrectable Errors in Storage Systems
The role of storage target allocation in applications' I/O performance with BeeGFS
Extracting and characterizing I/O behavior of HPC workloads

12:30 - 14:00

Lunch

14:00 - 14:30

Vendor Presentation: Min Li, Huawei
HS13
Compute 2030
Chair: Felix Wolf

14:30 - 15:30

Panel: Novel Hardware is Good, Programmable Hardware is Better
HS13
Novel Hardware is Good, Programmable Hardware is Better
Chair: Michèle Weiland

15:30 - 16:00

Coffee Break

16:00 - 17:00

Best Paper Nominees
HS13
Chair: Trilce Estrada
Improving Object Placement Methodology for Hybrid Memory Systems in HPC
Efficient Hierarchical State Vector Simulation of Quantum Circuits via Acyclic Graph Partitioning

18:00 - 23:00

Banquet
Guided tour (18:00 - 19:00)
Reception on castle terrace (19:00 - 20:00)
Dinner (20:00 - 23:00)

Conference - Friday, September 9

8:45-9:15

Registration

9:15 - 9:30

Award Ceremony and Cluster 2023 Presentation
HS13

9:30 - 10:30

Keynote: Rio Yokota, Tokyo Institute of Technology
HS13
Matrices in Deep Neural Networks and How to Compute Them in Parallel
Chair: Abhinav Bhatele

10:30 - 11:00

Coffee Break

11:00 - 12:30 -- Parallel Sessions

Operations and ML Training Strategies
HS13
Chair: Shadi Ibrahim
fairDMS: Rapid Model Training by Data and Model Reuse
ALBADross: Active Learning Based Anomaly Diagnosis for Production HPC Systems
HPC Storage Service Autotuning Using Variational-Autoencoder-Guided Asynchronous Bayesian Optimization

Node Technologies
HS10
Chair: Mohamed Hassan
Enabling Dynamic Virtual Frequency Scaling for Virtual Machines in the Cloud
SVAGC: Garbage Collection with a Scalable Virtual Address Swapping Technique
The Cost of Flexibility: Embedded versus Discrete Routers in CGRAs for HPC

12:30 - 14:00

Lunch

14:00 - 15:30 -- Parallel Sessions

Tensors & Linear Algebra
HS13
Chair: Trilce Estrada
BALA-CPD: BALanced and Asynchronous Distributed Tensor Decomposition
Optimizations of H-matrix-vector Multiplication for Modern Multi-core Processors
Optimizing Irregular-Shaped Matrix-Matrix Multiplication on Multi-Core DSPs

Distributed Memory Applications
HS10
Chair: Dirk Pleiter
Painless Transposition of Reproducible Distributed Environments with NixOS Compose
Integrating process, control-flow, and data resiliency layers using a hybrid Fenix/Kokkos approach
Fast Dynamic Updates and Dynamic SpGEMM on MPI-Distributed Graphs

15:30 - 16:00

Coffee Break

16:00 - 17:30 -- Parallel Sessions

Deep Learning
HS13
Chair: Rio Yokota
HVAC: Removing I/O Bottleneck for Large-Scale Deep Learning Applications
AutoPipe: A Fast Pipeline Parallelism Approach with Balanced Partitioning and Micro-batch Slicing
HPH: Hybrid Parallelism on Heterogeneous Clusters for Accelerating Large-scale DNNs Training

Shared Memory
HS10
Chair: Jay Lofstead
Recursive Multi-Section on the Fly: Shared-Memory Streaming Algorithms for Hierarchical Graph Partitioning and Process Mapping
MemGaze: Rapid & Effective Load-Level Memory Analysis
Spark Meets MPI: Towards High-Performance Communication Framework for Spark using MPI