Close

Session

Workshop: The 7th Annual Parallel Applications Workshop, Alternatives To MPI+X (PAW-ATM 2024)
DescriptionSupercomputers get faster and more complex every year. MPI, long the dominant model for distributed computation, has adapted by combining with models for intra-node parallelism (e.g., OpenMP, CUDA). These MPI+X hybrids offer performance but demand significant programmer effort to write, debug, and tune applications. Alternatives to MPI+X are worth exploring as programmer productivity becomes a major component of the time to science. Alternatives include parallel programing languages (e.g., Chapel, Regent, Fortran 2018), general purpose libraries (e.g., Charm++, COMPSs, HPX, Legion, UPC++), and domain specific libraries (e.g., Arkouda, Dask, Spark). With many options to choose from, it is hard for programmers to know which alternative models are appropriate for their application and for programming model developers to understand the opportunities for improvement. Through discussion of specific applications, PAW-ATM brings together application experts and programming model developers to improve applications and models.
Event TypeWorkshop
TimeSunday, 17 November 20249am - 5:30pm EST
LocationB306
Tags
Heterogeneous Computing
Parallel Programming Methods, Models, Languages and Environments
Task Parallelism
Registration Categories
W
Presentations
9:00am - 9:02am ESTThe 7th Annual Parallel Applications Workshop, Alternatives To MPI+X (PAW-ATM 2024)
9:02am - 9:45am ESTSurvey of Technologies for Developers of Parallel Applications — Task-Based and Scale-Free Computing
9:45am - 10:00am ESTResource Adaptivity at Task-Level
Author/Presenter
PAW-Abstract
10:00am - 10:30am ESTPAW-ATM — Morning Break
10:30am - 10:50am ESTSpeeding-Up LULESH on HPX: Useful Tricks and Lessons Learned using a Many-Task-Based Approach
PAW-Full
10:50am - 11:10am ESTLamellar: A Rust-based Asynchronous Tasking and PGAS Runtime for High Performance Computing
PAW-Full
11:10am - 11:30am ESTApplying a Task-Based Approach to Distributed Machine Learning Workflows
PAW-Full
11:30am - 11:50am ESTAccelerating Multi-GPU Embedding Retrieval with PGAS style Communication for Deep Learning Recommendation Systems
PAW-Full
11:50am - 12:10pm ESTMitigating synchronization bottlenecks in high-performance actor-model-based software
PAW-Full
12:10pm - 12:30pm ESTIntel® SHMEM: GPU-initiated OpenSHMEM using SYCL
PAW-Full
12:30pm - 2:00pm ESTPAW-ATM — Lunch Break
2:00pm - 2:45pm ESTA case study for using Chapel within the global aerospace industry
Session Chair
2:45pm - 3:00pm ESTExploring Suffix Array Algorithms in Chapel
Session Chair
PAW-Abstract
3:00pm - 3:30pm ESTPAW-ATM — Afternoon Break
3:30pm - 3:45pm ESTJust Write Fortran: Experiences with a Language-Based Alternative to MPI+X
Session Chair
PAW-Abstract
3:45pm - 5:30pm ESTAlternative programming models for applications at scale