The Denvr Difference
High Performance, Accelerated Compute - Low Environmental Impact
The Convergence of HPC and the Cloud
The next revolution in research computing is cloud transformation. Denvr Batch is based on Slurm Workload Manager and runs on client dedicated bare-metal or virtual machine instances.
Jobs are fault-tolerant and can be configured with user resource limits or for elastic bursting to scale out compute nodes.
Data Center-Scale Performance
By leveraging GPU-powered parallel processing, users can run advanced, large-scale application programs efficiently, reliably, and quickly.
Denvr uses all-InfiniBand clusters to provide scalability across multiple compute nodes, enabling parallel and synchronized compute cycles.
Denvr uses NVMe SSDs to provide cost-effective, ultra high-performance, scalable storage for compute workloads. GPU nodes are additionally equipped with local attached NVMe scratch files to ensure workloads are never delayed.
Combined with Denvr's InfiniBand HDR network, HPC jobs can expect over 10 GB/s throughput and sub-600 nanoseconds latency.
GPU-Accelerated Application Catalog
Users can deploy containers from the NVIDIA NGC™ hub which provides pre-built and tested AI, machine learning, and HPC frameworks. The catalog provides a range of options that meet the needs of data scientists, developers, and researchers with various levels of AI expertise.
NGC containers can be run on-prem, in the cloud, and edge or using hybrid and multi-cloud deployments.
Ready to get started?
Our sales team is here to help!