Cray is creating a production-ready, Arm-based supercomputer with the addition of Cavium ThunderX2 processors, based on 64-bit Armv8-A architecture, to the Cray XC50 supercomputer.
Cray is currently working with multiple supercomputing centers on the development of Arm-based supercomputing systems, including various labs in the U.S. Department of Energy and the GW4 alliance, a coalition of four research-intensive universities in the UK. Through an alliance with Cray and the Met Office in the UK, GW4 is designing and building “Isambard,” an Arm-based Cray XC50 supercomputer.
Cray XC50 supercomputers with the Cavium ThunderX2 processors will be available in both liquid-cooled cabinets and air-cooled cabinets to address various datacenter needs. Compute blades can be mixed and matched with Intel Xeon Scalable processors, Intel Xeon Phi processors, and NVIDIA Tesla GPU accelerators. The Arm-based Cray XC50 supercomputers will be available in the second quarter of 2018.
In related news, Cray offers artificial intelligence initiatives to advance deep learning for science and enterprise.
Cray is introducing:
- a new Cray Accel AI Lab to advance the development of deep learning technologies and workflows;
- new Cray Accel AI offerings, featuring NVIDIA Tesla V100 GPU accelerators, which provide a range of fast-start configurations leveraging the Cray CS-Storm accelerated cluster supercomputer along with a deep learning environment;
- a new Cray Urika-XC analytics software suite, including the addition of the TensorFlow open-source machine intelligence software library and support for Jupyter Notebooks; and
- an AI collaboration agreement with Intel, leveraging Intel’s AI technologies.
Cray is launching a deep learning innovation center–the Cray Accel AI Lab–to serve as a focal point for the application, incubation and advancement of supercomputing technologies in AI. The new lab will provide access to Cray systems powered by Intel Xeon Scalable Processors and NVIDIA Tesla GPU accelerators, and will bring together AI experts from Cray’s engineering teams, its customers and partners to provide educational content and classes in topics ranging from developer training to advanced optimization techniques of deep learning frameworks.
Cray is introducing Cray Accel AI offerings, fast-start configurations that range from a starter system ideal for AI exploration to a complete, production-level Cray cluster supercomputer for training and inference. Components include the Cray CS-Storm accelerated GPU system featuring NVIDIA Tesla V100 GPU accelerators powered by the NVIDIA Volta GPU architecture, and a comprehensive deep learning and analytics software environment from Bright Computing.
Intel and Cray have signed an agreement to collaborate on the development of a scalable software environment. The collaboration will deliver a productized software stack for deep learning at scale on Cray systems, along with targeted optimizations for deep learning and machine learning tools.
In more Cray news, Cray announces high performance computing (HPC) storage solutions including Cray View for ClusterStor; Cray ClusterStor L300N – a flash-based acceleration solution; and Cray DataWarp for the Cray XC50 supercomputer.
New features include:
- Cray ClusterStor L300N with NXD flash accelerator software;
- Cray DataWarp with transparent caching, which provides the ability to accelerate application performance; and
- Cray View for ClusterStor, which features the ability to trace and track the performance of Lustre.
The Cray ClusterStor L300N with NXD flash accelerator software optimizes multiple mixed workloads on the same storage platform.
Cray ClusterStor L300N, Cray View for ClusterStor, and DataWarp transparent caching software solutions are expected to be available in early 2018. Cray DataWarp for the Cray XC50 supercomputer is available now.
For more info, visit Cray.
Sources: Press materials received from the company.