Intel this week introduced that its processors, compute accelerators, and Optane DC persistent reminiscence modules will energy Aurora, the primary supercomputer within the US projected to characteristic a efficiency of 1 exaFLOP. The system is anticipated to be delivered in about two years, and goes past its preliminary Xeon Phi specification launched in 2014.
The US Department of Energy, Intel, and Cray have signed a contract below which the 2 corporations and DOE’s Argonne National Laboratory will develop and construct the Aurora supercomputer able to a “quintillion” floating level computations per second. The deal is valued at greater than $500 million, the system is anticipated to be delivered generally in 2021.
The Aurora machine shall be primarily based on Intel’s Xeon Scalable processors, the corporate’s upcoming compute accelerators primarily based on the Xe compute structure for datacenters, in addition to a next-generation Optane DC persistent reminiscence. The supercomputer will depend on Cray’s ‘Shasta’ structure that includes Cray’s Slingshot interconnect, that was introduced at Supercomputing again in November. The system shall be programmed utilizing Intel’s OneAPI and also will use the Shasta software program stack tailor-made for Intel.
Around two years in the past the DOE began its Exascale Computing Project to spur growth of {hardware}, software program, and purposes for exaFLOP-class supercomputers. The group awarded $258 million in analysis contracts to 6 expertise corporations, together with AMD, Cray, Hewlett Packard Enterprise, IBM, Intel, and NVIDIA. As it seems, Intel’s strategy was thought-about as probably the most environment friendly one for the nation’s first Exascale supercomputer.
It is noteworthy that ANL’s Aurora supercomputer again in 2014 was purported to be primarily based on Intel’s Xeon Phi codenamed Knights Hill produced utilizing the corporate 10 nm course of expertise. The plan modified in 2017, when Intel canned the Knights Hill in favor of a extra superior structure (and the truth that its Xeon processors had been approaching a Xeon Phi-like implementation). Apparently, Intel and its companions are assured within the new chips to proceed with the venture now.
The Aurora supercomputer will be capable to deal with each AI and conventional HPC workloads. At current, Argonne National Laboratory says that amongst different issues this machine shall be used for most cancers analysis, cosmological simulations, local weather modeling, discovering drug response, and exploring numerous new supplies.
“There is tremendous scientific benefit to our nation that comes from collaborations like this one with the Department of Energy, Argonne National Laboratory, industry partners Intel and Cray and our close association with the University of Chicago,” stated Argonne National Laboratory Director, Paul Kearns. “Argonne’s Aurora system is built for next-generation artificial intelligence and will accelerate scientific discovery by combining high-performance computing and artificial intelligence to address real world problems, such as improving extreme weather forecasting, accelerating medical treatments, mapping the human brain, developing new materials and further understanding the universe — and those are just the beginning.”
Sources: Intel, Intel, Argonne National Laboratory
