Cray will bring a Docker solution optimized for high-performance computing environments to its line of XC supercomputers starting in December and expanding into next year.
The application distribution technology announced this week will enable XC supercomputer customers to grow the number of applications that are available to them while reducing the time needed to deploy the applications on the system. The Docker solution will offer containers that hold the software stack—such as the application codes, shared libraries, base operating system files, user environment variables and dependencies—needed for an application to run better across multiple platforms, according to Cray officials.
“The supercomputing community continues to evolve in our shared quest for discovery and scientific breakthroughs,” Ryan Waite, senior vice president of products at Cray, said in a statement. “We are seeing an increasing number of developers using new technologies to solve their problems. For some of our customers, the use of languages like Python and R and system technologies like Docker is ideal for their modern supercomputing applications.”
The new Docker feature will be available to all Cray XC customers starting next month, officials said. It will be available on the CS400, Cray XE and Cray XK platforms in 2016.
The supercomputer company developed the technology through a project with one of its customers, the National Energy Research Scientific Computing Center (NERSC). The two organizations worked on a container product code-named Shifter, which extended Docker capabilities to high-performance computing (HPC) environments and Cray XC30 system users, officials said. The technology developed through the project was validated on an XC30 system at the NERSC dubbed Edison, and led to the creation of the Docker solution for large-scale computing environments.
“Container computing, exemplified by Docker, is poised to revolutionize the way scientific and technical computing is carried out,” Shane Canon, project engineer for advanced systems technology at NERSC, said in a statement. “Having these capabilities on HPC platforms, like those from Cray, will lower the barriers for many workloads and will boost productivity. Containers can also help with the process of science itself, since it can ease reproducibility and simplify sharing.”