HPC in the cloud service for Engineering
Quickly tap into the power of Sabalcore’s purpose-built high performance computing systems to crunch numbers and speedup simulation time. Increase production and lower over-all costs by reducing time-to-market. Sabalcore’s HPC Cloud is designed to run parallel applications quickly and reliably.
Parallel applications which use MPI rely heavily on distributing the domain process across many compute nodes in order to reduce the wall-time necessary to complete the simulation. Therefore, the network connection between the nodes is critical to the performance, scalability, and reliability of the process. Sabalcore utilizes the Nvidia Quantum Infiniband network platform which provides the ultra-low latency, full-bandwidth, connectivity between compute nodes necessary for parallel performance.
- High performance NVIDIA Infiniband interconnect platform
- High performance parallel filesystems for demanding I/O
- Direct-attached storage for high-throughput direct-to-disk I/O
- Bare-metal compute nodes running Linux OS for highest available performance per CPU cycle; better than vCPUs
- CLI and Remote Desktop capabilities
- Proprietary file transfer for fast data download/upload to the Sabalcore cloud