An Inside Look at the Magma Supercomputer
Earlier this year, Penguin Computing™, in partnership with Intel and CoolIT, deployed a new supercomputer to our customer, Lawrence Livermore National Laboratory. This system is the latest addition to the Commodity Technology Systems (CTS-1) contract, and is one of the first deployments of Intel® Xeon® Scalable 9200 series processors at this scale.
Magma, a 4.6 petaFLOPS supercomputer, is built on the Penguin Computing® Relion® 2142eAP server platform which consists of 752 nodes with Intel Xeon Platinum 9242 (Cascade Lake-AP) processors. The cluster also has 293 terabytes of memory, liquid cooling provided by CoolIT Systems. Like other CTS-1 systems, the fabric is based on Intel Omni-Path. Due to the higher performance node requirements, LLNL doubled the on-node network performance by adding a second Intel OPA host adapter for each node. Magma was built to support NNSA’s Life Extension Program and efforts critical to ensuring the safety, security and reliability of the nation’s nuclear weapons in the absence of underground testing.“The quick collaboration amongst all the stakeholders allowed for fast design, contract execution, delivery, and ultimate acceptance of Magma,” added Gudenrath. “Our final goal was achieved when we completed several initial high-performance Linpack runs and submitted these for qualifying on the November 2019 Top500 list.” After initial benchmark tests and validations in the Penguin Computing facility, Magma placed 69th on the Top500 list.
To learn more about Magma, it’s capabilities, and the process of building and deploying this cutting edge supercomputer, read this case study, developed by our partners at Intel.[mk_padding_divider size=”10″][mk_button_gradient size=”medium” grandient_color_from=”#005b9a” grandient_color_to=”#004884″ grandient_color_fallback=”#005b9a” url=”https://www.penguinsolutions.com/computing/wp-content/uploads/2020/05/penguin-computing-lawrence-livermore-magma-case-study.pdf” target=”_blank”]GET THE CASE STUDY[/mk_button_gradient][mk_padding_divider size=”20″]