Skip to content Skip to navigation

NEXTGenIO partners Intel and Fujitsu are developing a prototype system based around Intel’s Optane DC Persistent Memory™ next-generation non-volatile memory technology.


NEXTGenIO partners Intel and Fujitsu are developing a prototype system based on Intel’s Optane DC Persistent Memory™ next-generation non-volatile memory technology.

The prototype system will be used to explore how to make best use of this new technology in the world of I/O intensive high-performance scientific computing. In these initial stages of the project, all partners are working together on defining the architecture of the prototype system, with the aim of creating a platform that will be transformational not only for HPC, but across the entire computing spectrum.


The project's major goal is to create a new system based on the new memory technlogy. This requires designing new architectures that can effectively take advantage of the technology.

The first step in this processhas been to capture the requirements of the system we are developing. We can then use these requirements to define the architectures for the different components in the prototype we will eventually build. In total, there are three different architectures that the project will need to develop:

  • The hardware architecture: this will define the hardware components required, and how they will be connected and packaged. This is more complex than simply defining the compute nodes for our prototype, as it must cover any login nodes, management resources, networks, and I/O systems.
  • The software architecture: This is one of the key goals of the project. We will define a complete software architecture for enabling the use of NVRAM in scientific applications. The project will utilise existing software components that are already under development for NVRAM usage, as well as implementing any new software solutions that will be needed to effectively use this technology.
  • The data architecture: This will provide users with an understanding of how data can move and when it can be located in our hardware and software architectures, enabling users and developers to understand how to optimise performance on the system for their applications.

The initial steps of the process were to capture all the requirements (from the application, development, and system deployment, operation, and management perspectives) and the constraints (from the hardware and software we know we will be using) that will be relevant to the prototype system.  We defined a number of different usage scenarios, first by imagining ways a HPC system with NVRAM technology may be deployed, and then refined these into distinct use cases.  The use cases were then examined to generate specific requirements for the prototype system. Potential constraints were gathered from the project's hardware partners, Fujitsu and Intel, by examining the documentation for their hardware components and using their experiences of designing and creating systems to create a realistic set of constraints likely to affect the prototype.