Overview

The notion of cloud computing has changed the way how we utilize computing resources. Since High-Performance Computing (HPC) has long been suffered from under- or over-utilization of resources, many HPC researchers are trying to adapt HPC applications to the cloud environment. With proper adaptation, HPC applications are able to enhance their resource utilization ratio and scalability by using virtualized and on-demand resources on clouds. While we discuss HPC on clouds, we should discuss the parallel programming models as well. Various parallel programming models and their frameworks (e.g., MPI, OpenMP, OpenCL, CUDA, and MapReduce) has been proposed for parallel computing. For example, the MapReduce programming model has been used for various big data processing applications since it helps to reduce the complexity of problem parallelization such as decomposition, communication, and scheduling. However, a parallel programming model should be carefully selected for HPC applications to achieve high-performance and efficient resource usage because their target hardware architectures (e.g., many-core, GPU, interconnect, etc.) are different as well as the abstraction levels. For example, MapReduce may not be a suitable selection of parallel programming model for a large-scale graph data processing problem. In addition, since traditional parallel programming models, such as MPI, are implemented for a single tenant cluster environment, applying these models to HPC applications on the cloud is a challenging in terms of resource management.


Submitting a Paper

The 2nd International Workshop on Parallel Programming Models in High-Performance Cloud (ParaMo 2020) will provide a venue for researchers to discuss recent results and the future challenges to parallel programming models in high-performance cloud. The topics include, but are not limited to:
- Parallel programming models for large scale data processing (e.g., MapReduce) in the cloud
- Parallel programming models for massively parallel computing (e.g., MPI, OpenMP, and OpenCL) in the cloud
- High-performance networking for parallel programming models in the cloud
- High-performance storage for parallel programming models in the cloud
- Heterogeneous resource management (e.g., many-core and GPU) for parallel
- programming models in the cloud
- Load balancing schemes for HPC applications in the cloud
- Runtime support for parallel programming models in the cloud
- Energy efficient resource management and parallel programming models in the cloud
- Resource management for virtualized environments
- Performance evaluation for parallel applications in the cloud
- Configurational optimization for parallel applications in the cloud

The submissions should follow the LNCS format. They should be between 10 to 12
pages. Each submission will be reviewed by at least three members of program
committee, on the basis of relevance, originality, and clarity. Paper should be
submitted electronically via EasyChair.

Journal Special Issue
The papers accepted for ParaMo 2020 will be invited to submit an extended version to a special issue of Wiley's Concurrency and Computation: Practice and Experience journal (confirmed).

Please check our website for periodic updates:
https://sites.google.com/site/paramoworkshop2020/

Workshop Organizers

Program Co-Chairs
Hyun-Wook Jin (Konkuk Univ., Korea)
Sangyoon Oh (Ajou Univ., Korea)

Advisory Committee
Geoffrey C. Fox (Indiana Univ., USA)
Dhabaleswar K. Panda (Ohio State Univ., USA)

Publicity Chair
Yin-Goo Yim (Konukuk Univ., Korea)

Program Committee
Seung-Hee Bae (Intel, USA)
Jee Choi (Univ. of Oregon, USA)
Jong Choi (Oak Ridge National Lab., USA)
Cheol-Ho Hong (Chung-Ang Univ., Korea)
Xiaoyi Lu (Ohio State Univ., USA)
Blesson Varghese (Queen's Univ. Belfast, UK)
Wenjun Wu (Beihang Univ., China)
Beytullah Yildiz (Atilim University, Turkey)
Weikuan Yu (Florida State Univ., USA)