Advanced Research Computing | Capital Area Advanced Research and Education Network (CAAREN)

CAAREN (Capital Area Advanced Research and Education Network) supports advanced research computing by providing robust infrastructure, cutting-edge technology, and expert guidance, enabling researchers to tackle their most demanding computational challenges efficiently and effectively.

Among the key computing models supported by CAAREN are High Throughput Computing (HTC) and High Performance Computing (HPC), each with distinct purposes.

HTC excels in handling numerous small, independent tasks, making it ideal for extensive data analysis and research.
HPC is designed for large-scale, complex computations, crucial for simulations and high-speed data processing. Together, these models empower researchers to push the boundaries of innovation and discovery.

Focus: Maximizing computational speed and power.
Tasks: Large, complex, interdependent tasks.
Example Usage: Weather forecasting, molecular dynamics, engineering simulations.
Infrastructure: Supercomputers with high-speed interconnects.
Advantages: Faster computation, scalability, parallel processing

Focus: Maximizing the number of tasks completed over time.
Tasks: Many small, independent tasks.
Example Usage: Scientific research, data analysis, simulations.
Infrastructure: Distributed computing clusters.
Advantages: Cost-effective, flexible, reliable

High Performance Computing (HPC)

Explore the high performance computing solutions provided by CAAREN to support research that relies on data analytics and supercomputing applications.

Benefits & Specifications

Diverse High-Performance Computing Clusters: GW offers three distinct clusters - Pegasus, Cerberus, and a high throughput computing cluster, catering to various research needs that require high-performance computing for data analysis.
Professional Management and Support: All clusters are professionally managed by GW Information Technology, with dedicated computational staff from various academic departments, ensuring smooth operation and technical support.
Open Access to University Community: The HPC resources are accessible to the entire university community, supporting research, academic classroom use, and student projects.
Advanced Computing Capabilities: The new Pegasus cluster offers advanced computing capabilities with 210 compute nodes, GPU, high memory, and high throughput nodes. It utilizes CentOS 7.4 and the SLURM job scheduler, capable of achieving 2.14 PFLOPs Single Precision.
Innovative High Throughput Cluster: The Condor cluster, rooted in Open Science Grid (OSG), represents an innovative approach to high throughput computing. It’s under continuous development, with a shift towards OpenHPC’s cluster management tools and Ansible for enhanced performance and efficiency.