CS-Storm GPU Supercomputer

Accelerated computing system consisting of multiple high-density multi-GPU server nodes

Contact Us

Cray® CS-Storm™ Accelerated GPU Cluster System

Cray® CS-Storm™ cluster supercomputers tackle the toughest extreme HPC and artificial intelligence (AI) workloads. Designed for speed, architected for scale and integrated for production use, the Cray CS-Storm GPU supercomputer is your path to exploiting the performance available from the latest NVIDIA® Tesla® GPUs.

Designed for Speed
Powered by NVIDIA® Volta or NVIDIA® Pascal™ architecture GPUs, each CS-Storm server node and rack system is integrated by Cray to deliver maximum accelerated computing performance across the broadest range of HPC and AI uses.

Architected for Scale
While others focus on small cluster nodes, Cray architects and delivers accelerated computing systems built for production-level HPC and AI applications that require scaling beyond the accelerator or node. Whether it’s deep learning training, signal processing, reservoir simulation, high-performance microscopy or medical image processing, the Cray CS-Storm system is architected with scaling in mind.

Integrated for Production Use
Designing, implementing and using a GPU cluster system doesn’t have to be difficult. You can rely on Cray, the expert in production supercomputing, to simplify your environment. Our accelerated computing systems are built using open, industry-standard technologies. We deliver a complete solution for HPC and AI including high-performance networking, storage, a comprehensive software environment for cluster management and developer productivity, and the latest machine and deep learning frameworks.

Cray CS-Storm 500GT System

The Cray CS-Storm 500GT configuration scales up to 10 NVIDIA Tesla Volta V100, Pascal P40 or P100 GPUs or Nallatech FPGAs by leveraging the flexibility and economics of PCI Express as well as the power of today’s advanced Intel® Xeon® processor Scalable family.

Cray CS-Storm 500NX System

The Cray CS-Storm 500NX configuration scales up to eight NVIDIA Tesla Volta or Pascal architecture GPUs (V100, P100) using NVIDIA® NVLink™ to reduce latency and increase bandwidth between GPU-to-GPU communications, enabling larger models and faster results for AI and deep learning neural network training.

Featured Resources

Cray Accel AI Deep Learning Solutions

We give you the tools and support to guide you to success whether you’re just starting out or ready for production AI.

Cray CS-Storm GPU-Accelerated Cluster Supercomputer Brochure

Purpose-built for the most demanding machine learning workloads, Cray CS-Storm technology provides customers with a powerful, accelerator-optimized solution for moving AI applications from pilot to production.

Processors and Accelerators

CS-Storm 500GT: Up to eight 450W or ten 400W accelerators; NVIDIA® Tesla® V100, P40 or P100 PCIe GPU accelerators supporting up to 32 GB of memory; Nallatech FPGA accelerators; two Intel® Xeon® Scalable processors; and PCIe in a balanced configuration

CS-Storm 500NX: Up to eight NVIDIA Tesla V100 or P100 SXM2 GPU accelerators supporting up to 32 GB of memory, two Intel® Xeon® E5-2600 v4 processors, and NVIDIA® NVLink™ for 8-way GPU-to-GPU communications


InfiniBand™ or Intel® OmniPath architecture


CS-Storm 500GT: 19" 3U or 4U rackmount chassis, 16 hot-swap 2.5” drives (up to 8 NVMe)

CS-Storm 500NX: 19" 4U rackmount chassis, 12 2.5" drives (up to 4 NVMe)

AI Software

Supports Bright Cluster Manager™ for HPC and an optional Bright Deep Learning extension

Machine learning frameworks including Caffe, Chainer, CNTK, Keras, MXet, Pytorch, Scikit-learn, Tensorflow, Theano, Torch, Dynet, Intel Neon, TensorFlowOnSpark, BigDL, Deeplearn4j

Libraries/tools: basel, CUB, CUDA, CUDNN, NVIDIA® DIGITS, flags, glob, DHF5, lmdb, NCCL, OpenCV, Protobuf, Theano, Torch