HPE Solutions for Deep Learning

Print Friendly, PDF & Email

This is the fifth article in a series taken from The inside HPC Guide to The Industrialization of Deep Learning

Apollo 6500 / 6000 series:

Coming in the second half of 2016: The HPE Apollo 6500 System provides the tools and the confidence to deliver high performance computing (HPC) innovation. The system consists of three key elements: The HPE ProLiant XL270 Gen9 Server tray, the HPE Apollo 6500 Chassis, and the HPE Apollo 6000 Power Shelf. Although final configurations and performance are not yet available, the system appears capable of delivering over 40 teraflop/s double precision, and significantly more in single or half precision modes. High-bandwidth, low-latency networking is tightly coupled to the accelerators to take full advantage of network capability. Two x16 PCIe Gen3 slots support a choice of high speed fabrics including InfiniBand, OmniPath or Ethernet. In combination with the modular Apollo 6000 rack system which is capable of supporting up to 20kW per rack has the potential to deliver a very impressive industrial grade production environment.


HPE Cognitive Computing Toolkit / Use cases:
HPE CCT is a functional programming language, optimizer, and distributed runtime for ultra-scale real time data analytics that delivers a simplified programming model that frees the data scientist from the challenges of parallel and GPU programming with demonstrated scaling across 50K GPU cores.
High level Diagram of the Initial Release of the HPE Cognitive Computing Toolkit.


HPE CCT brings some currently unique capabilities to the deep learning development market in the ability to write kernels and optimize kernels in a high level way that eliminates the need for dropping into a second language. The compute models for CCT and TensorFlow are similar which provides a simple to use python API to generate CUDA and C code that can be passed to TensorFlow to generate custom operations in a much easier way delivering both higher productivity and higher performance.


Support for embedding in native applications is planned for subsequent releases together with further open source community activities.

This series has explored The Industrialization of Deep Learning.

If you prefer you can download the complete inside HPC Guide to The Industrialization of Deep Learning courtesy of Hewlett Packard Enterprise