Ten Exascale supercomputers by 2023

The US is spending $1.8 billion on three exascale supercomputers which should be completed in 2021-2022. The US is building the A21 at Argonne national labs, Frontier at Oak Ridge national lab and El Capitan at Livermore national lab.

By 2022, the fastest of US exaflop supercomputers could be 1.3 exaflops or more.

China has built three exascale supercomputer prototypes.

China has launched the prototype of Shuguang, an exascale supercomputer being developed by Dawning Information Industry, also known as Sugon.

Shuguang machine is will be operated in national supercomputing centers in Shanghai and Shenzhen. The first two exascale prototypes were announced during the summer of 2018.

The first was built at the National Supercomputing Center in Tianjin. It is the prototype for the Tianhe-3 exascale machine. It is presumed to be the Arm-based supercomputer, based on Phytium’s Xiaomi platform.

The second prototype is the precursor to the Sunway exascale system, which is set to be installed at the National Supercomputing Center in Jinan. The Sunway machine is expected to be based on a future version of the ShenWei processor, the latest version of which is used to power the 93-petaflop Sunway Taihulight supercomputer. No details were revealed about the nature of the processors powering the two prototypes.

The first exaflop supercomputer will likely be the Tianhe-3 in 2020. Nextbigfuture believes that China will push ahead and make exascale supercomputers based upon each of the prototypes.

China is investing $470 million in the first exaflop supercomputer.

Tianhe-2A, also known as Milky Way-2A, is now the fourth fastest supercomputer. It had a major upgrade that replaced its five-year-old Xeon Phi accelerators with custom-built Matrix-2000 coprocessors. The new hardware increased the system’s HPL performance from 33.9 petaflops to 61.4 petaflops, while bumping up its power consumption by less than four percent.

Japan and Europe also have billion-dollar efforts to develop exascale computing

Japan is also aiming for an exascale machine with a successor to its K supercomputer at the RIKEN Advanced Institute for Computational Science in Kobe, though the head of the project has said that its delivery could slip to 2021 or 2022. The European Union could cross the exascale threshold in 2021, according to market research firm Hyperion Research.

But that won’t be the end of the race. The four supercomputing powers are convinced they need to push the frontiers in order to compete in a wide range of scientific disciplines, defense technology, industrial technology, and computer products.

Bottom of the Top 500 are currently one or two Petaflop supercomputers

The bulk of the supercomputers on the top 500 list are systems with configuration similar to this example ranked at 307.

18 AGX-2 servers, each outfitted with eight of NVIDIA’s new V100 GPUs and two Intel Skylake processors. The 144 GPUs will deliver over one peak petaflops of double precision computing and more than 17 petaflops of deep learning performance, the latter courtesy of the V100’s specialized Tensor Cores. The servers will be hooked together with Mellanox EDR InfiniBand.

Lenova, HP Enterprise, Inspur, Sugon and Cray make 70% of the top 500 supercomputers.

Lenovo now has an upgrade SD650 node. Germany is putting 6500 of those to make a 26 petaFLOP supercomputer.

With air cooling, a data center consumes about 60% of the server power to cool the servers,
* 40% of server power with chilled water cooling,
* less than 10% with warm water like the SD650.

The potential savings for a supercomputer cluster that has a power consumption of 4-5 MW can be €100,000’s savings per year over 4-5 years.

The supercomputers in the 200-500 spot should shift from one to two petaflops to five to ten petaflops by 2020 and then to ten to thirty petaflops by 2023.

DUG immersive dielectric cooling

A 250 petaflop supercomputer immersed in liquid coolant is being built in Texas. In 2019, it could be the world’s most powerful supercomputer.

40,000 server immersed in coolant are going into a data center in Houston. It is being built by DownUnder GeoSolutions (DUG).

It will perform cutting-edge computer modeling for energy companies and bring new levels of precision to oil and gas exploration. It will be housed in the Skybox Datacenters facility in Houston’s Energy Corridor, where DUG has leased 15 megawatts of capacity. The deal, represented by Bennett Data Center Solutions, is the largest colocation transaction in Houston’s history.

720 enclosures using the DUG Cool liquid cooling system, which fully submerges servers in tanks filled with dielectric fluid. This will reduce the huge system’s energy usage by about 45 percent compared to traditional air cooling.

A 60-megawatt power system could be scaled for an exaflop supercomputer in 2020 or 2021 using the Australian technology.

Immersive Dielectric Cooling

The cooling system fully submerges standard high-performance computing (HPC) servers into specially-designed tanks that are filled with polyalphaolefin dielectric fluid. The fluid is non-toxic, non-flammable, biodegradable, non-polar, has low viscosity, and most importantly, doesn’t conduct electricity. The unique part of this design is that the heat exchangers are very simple and submerged with the computer equipment, meaning that no dielectric fluid ever leaves the tank. A water loop runs through the rooms and to each heat exchanger.

The dielectric fluid is cooled and circulated around the extremely hot components in the compute servers. This innovative oil-cooling solution has high thermal capabilities and a large operating temperature range.

Traditionally, companies like DUG spend half of their revenue on compute centres. Energy use can be a significant and crippling expense in a conventional data centre. With DUG Cool, the thermal qualities of the fluid mean that condensed-water chillers can be used rather than refrigeration, saving 25-30% of total power usage. Removing all server fans, which aren’t needed in a fluid-immersion system, reduces power consumption by a further 20%. That’s a total power saving of at least 45%.

The over 1000x thermal capacity of the fluid vs. air means that components never get hot, reducing their mean time to failure. Fluid-immersed computers fail at a much lower rate, considerably reducing maintenance costs and expensive down-time.

Better cooling reduces the operating costs which enables more budget for faster overall systems.

DownUnder GeoSolutions

DownUnder GeoSolutions (DUG) is an Australian-owned high-tech geosciences company.