Top 500 Supercomputers: Intel Xeon Cascade Lake Nearly Triples Sites

While Intel's Xeon Cascade Lake processors triple in the world's top supercomputers, AMD notches a few victories with the first top supercomputers using its EPYC processors in the new Top500 list. Meanwhile, the relevance of Nvidia's GPU accelerators continues to grow.

Intel tripled the number of the world's top supercomputers using its second-generation Xeon Scalable processors, code-named Cascade Lake, while its top rival, AMD, saw its EPYC processors enter the fray for the first time with four leading high-performance computing systems.

That's according to the latest list from Top500, the organization that has been tracking the world's top 500 supercomputers annually since 1993. The fall 2019 list was published on Monday, the first day of the Supercomputing conference happening this week in Denver.

[Related: Intel Unveils 7nm Xe Ponte Vecchio GPU For HPC, AI Workloads]

In addition to CPU gains made by Intel and AMD, Nvidia saw the number of supercomputers using its GPU accelerators grow to 136 from 127 in June, with the number of systems using Nvidia's top data center GPU, the Tesla V100, reaching 90. AMD, for its part, saw its Radeon Instinct MI60 accelerator enter the list with one system.

The new Top500 list comes as Intel, AMD and Nvidia continue to battle over emerging high-performance computing and artificial intelligence workloads that are increasingly becoming relevant to commercial organizations and the enterprise at large. The list also teases the increasing relevance of Arm-based processors as the British semiconductor company pushes its Neoverse CPU platform.

"It's really about the start of an entirely new era of computing, where exascale technologies will be needed to run new digital transformation workloads," Peter Ungaro, CEO of HPE-owned supercomputing vendor Cray, said ahead of the Supercomputing conference.

Top500 said the United States and China continued to dominate the list, with the latter country notching its total number of supercomputers to 227, more than half of the world's top 500 supercomputers. Meanwhile, the U.S. continued to lead in performance with 37.8 percent share of the list's aggregate performance, ahead of China's 31.9 percent performance share.

The list shows that the total aggregate performance of the world's top 500 supercomputers raised to 1.14 petaflops from 1.02 petaflops from the Top500 list released in June. A petaflop amounts to one quadrillion floating-point calculations per second.

The number of supercomputers using Intel's Xeon Cascade Lake processors nearly tripled to 26 in the fall 2019 list from nine in the June 2019 list, when the server processors debuted in the top 500. The semiconductor giant continued to hold dominance over the top 500 overall, representing 470 systems, or 94 percent, split between multiple generations of Xeon and Xeon Phi processors.

The fastest supercomputer running Cascade Lake is Frontera, an academic supercomputer at the Texas Advanced Computer Center at the University of Texas. Ranked No. 5, the supercomputer cluster consists of Dell EMC PowerEdge C6420 servers, 28-core Intel Xeon Platinum 8270 processors and Mellanox InfiniBand HDR interconnect fabric, representing a total of 448,448 CPU cores.

Out of the systems running Cascade Lake, two use one of Intel's top-of-the-line SKUs from the processor lineup, the Xeon Platinum 9242, which comes with 48 cores, a base frequency of 2.3 GHz and a turbo frequency of up to 3.8 GHz. Unlike the rest of the Cascade Lake lineup, the Xeon 9200 processors are not standalone parts, meaning they must be purchased as part of Intel's 9200WK compute platform.

Meanwhile, AMD notched its first victories in the top 500 list for its EPYC server processors that debuted in 2017 as direct competition to Intel Xeon. Four systems total use AMD EPYC, split evenly between the chipmaker's first-generation Naples processor and second-generation Rome processor. Two other supercomputers use AMD-based processors: one for AMD's now-defunct Opteron processor and another for the Higon Dhyana, a processor developed by AMD's joint venture in China.

The top supercomputer using AMD-based processors is the Advanced Computing System for Sugon, a state-owned supercomputer developer in China that is involved in the AMD joint venture, Higon, which licensed x86 processor technology from AMD to build the Dhyana processor. Sugon, Higon and Sugon-owned THATIC, which is also involved in the joint venture, were among the Chinese companies blacklisted from buying U.S. technologies by the U.S. Department of Commerce in June.

Sugon's Advanced Computing System, ranked No. 48 on the list, consists of Sugon TC8600 servers, 32-core Higon Dhyana CPUs, an unnamed deep computing accelerator processor and a 6D-Torus interconnect, representing a total of 163,840 CPU cores and 131,072 accelerator cores,

The top supercomputer running AMD EPYC Rome processors, on the other hand, is the Joliot-Curie research supercomputer in France. Ranked at No. 59, the supercomputing cluster consists of Atos' BullSequana XH2000 servers, 64-core AMD EPYC 7H12 processors and Mellanox InfiniBand HDR interconnect fabric, representing a total of 160,000 CPU cores.

Meanwhile, IBM continued to maintain a small share of supercomputers running with its Power processors at a total of 14, down one from the June list. And the number of supercomputers using Arm-based processors doubled to two systems from the June top 500 list with Fujitsu's A64FX prototype, which uses the Japanese company's new 48-core Fujitsu A64FX Arm processors.