Search

HPC clouds may provide fast access to fully configurable and dynamically scalable virtualized HPC clusters to address the complex and challenging computation and storage-intensive requirements. The complex environmental, software, and hardware requirements and dependencies on such systems make it challenging to carry out our large-scale simulations, prediction systems, and other data and compute-intensive workloads over the cloud. The article aims to present an architecture that enables HPC workloads to be serverless over the cloud (Shoc), one of the most critical cloud capabilities for HPC workloads. On one hand, Shoc utilizes the abstraction power of container technologies like Singularity and Docker, combined with the scheduling and resource management capabilities of Kubernetes. On the other hand, Shoc allows running any CPU-intensive and data-intensive workloads in the cloud without needing to manage HPC infrastructure, complex software, and hardware environment deployments.

The significance of earth observation data spans diverse fields and domains, driving the need for efficient management. Nevertheless, the exponential increase in data volume brings new challenges that complicate processing and storing data. This article proposes an optimized multi-modular service for earth observation data management in response to these challenges. The suggested approach focuses on choosing the optimal configurations for the storage and processing layers to improve the performance and cost-effectiveness of managing data. By employing the recommended optimized strategies, earth observation data can be managed more effectively, resulting in fast data processing and reduced costs. [ABSTRACT FROM AUTHOR]

Kn¨odel graphs of even order n and degree 1 ≤ δ ≤ log2(n),Wδ,n, are regular graphs that have an underlying topology that is time optimal for algorithms gossiping among n nodes. Because of their distinctive properties, Kn¨odel graphs act as a time-optimal topology for broadcasting and gossiping, thus arising in many settings, including social and communication networks or agent-based modeling simulations. Experimentation, often based on the extensive generation and analysis of complex networks, relies on high-performance computational resources to efficiently simulate the flow of information. The efficacy of such processing commonly depends on parallel processing and proper provisioning of distributed resources. This article aims to introduce a runtime in the partitioned global address space model that is optimized for performance and designed to improve the processing of Kn¨odel graphs. The sequential and parallel generation of synthetic datasets, and simulation of push-based, and broadcast-based gossiping algorithms with detailed analysis of resource usage and runtime have been studied. [ABSTRACT FROM AUTHOR]

The parallel scaling of GROMACS package has been studied in detail by estimation the performance of machine in presence of new generation graphical cards. This paper aims to estimate the speedup and performance issues depending on GROMACS network parametrics and as well as the version. The obtained results show that depending on threads/MPI ratio the new version of GROMACS plays rather good than previous one, meanwhile there are cases when the performance values are nearly same. The presence of GPU is significantly increasing the performance scaling up to ten time acceleration. Pinning threads to cores guaranties up to 15-35% increase. [ABSTRACT FROM AUTHOR]