The Supercomputer Race After ExaFlops is Super Cool

The first Exaflop supercomputers are just scaling up with energy efficient modifications the technologies that have enabled the 100-200 petaFlop supercomputers.

Super Cool Technologies Beyond the Exaflop Supercomputer

DownUnder GeoSolutions (DUG) has started making petaflop supercomputers that are much more energy efficient by switching to dielectric liquids for cooling. They are cutting energy usage and operating costs in half. Early in 2019, they will activate the 250 PetaFlop (single precision) DUG Cloud computing system.

Race for Superconducting Supercomputers

In 2018, China invested $145-million into a five-year catch-up effort to fabricate their own superconducting computers by 2022.

Superconducting Josephson junctions switch quickly (~1 picosecond), dissipate little energy per switch (< 10^-19 Joules), and communicate information via small current pulses that propagate over superconducting transmission lines nearly without loss. Significant technical obstacles have been overcome which prevented serious exploration of superconducting computing. The foundations have been set for a major breakthrough. Superconducting supercomputers may be capable of 1 PFLOP/s for about 25 kW and 100 PFLOP/s for about 200 kW, including the cryogenic cooler. IARPA began funding superconducting computers as a long-term solution to the power-cooling problem of supercomputers. Cryogenics make superconducting chips equipped with Josephson junctions 1,000 times more expensive to cool than CMOS, but they are 100,000 times more energy efficient in operation, yielding an 100-fold increase in overall energy efficiency. Superconducting computers could enable clock speeds that 150 times faster. Current CMOS is limited to 5GHz limit for CMOS but superconductors could reach 770GHz. IARPA started two superconducting programs—one to advance superconductor electronic design automation (EDA) called SuperTools, and a separate superconductor-to-digital computer interface effort called SuperCable. SuperTools is developing a superconducting circuit design flow with a comprehensive set of Electronic Design Automation (EDA) and Technology Computer Aided Design (TCAD) tools for Very-Large-Scale Integration (VLSI) design of Superconducting Electronics (SCE). The University of Southern California and Synopsys are working on SuperTools. Mikhail Dorojevets, a Google Scholar and professor at Stony Brook University, has criticized superconducting computers. The maximum superconducting clock speed and the current chips do not represent the potential speeds or representative circuits that can be produced.

SuperCables

IARPA funded technology and techniques for energy-efficient, high data rate transmission of digital signals between computing systems operating at room and cryogenic temperatures. The SuperCables program is trying to demonstrate components to convert from low-level electrical signals in circuits operating at a temperature of approximately 4 kelvins to conventional optical signals at room temperature. Pending results of this program, IARPA may support a follow-on program to develop the complete system for data transmission between room temperature and 4 kelvins.

D-Wave System Has Made Progress on Some of the Superconducting Computing Chalenges

D-Wave Systems has been using cryogenic superconducting chips for quantum annealing systems. They have pushed forward technologies that are needed for superconducting supercomputers. They have fabricated the smallest superconductor chip feature sizes today (240 nanometers, fabricated by D-Wave’s foundry, SkyWater Technology). They have successfully adapted existing EDA tools to superconductors as SuperTools aims to do with Synopsys’ EDA tools. They have solved the problem of interfacing superconducting circuitry with CMOS but this was done at a smaller scale than IARPA’s SuperCable goal.

Other Non-Cryo Technology to get to Many ExaFlops

There are other technologies which could enable supercomputers beyond Exaflop speeds.

IBM is developing AI focused 8 bit technology that will be 200-400 times faster than a Google third generation TPU (Tensor Processing Unit). 1000 TPUs have been combined into a pod which can achieve 100 Petaflop (16 bit) processing speed. If the scaling issues can be resolved, then IBM analog and digital AI focused systems or Google technology could achieve 1000 ExaFlops or a ZettaFlop around 2030.

Optalysis is still funded to develop multi-exaflop optical supercomputing systems. They are targeting around 2022 for something around 20 ExaFlops.

20 thoughts on “The Supercomputer Race After ExaFlops is Super Cool”

  1. France is determined, but what about the other 27 members or 26 depending on Brexit? Arianespace was always a French project, in fact Britain even dropped its Black Arrow launcher in the early 1970s before joining the EU in order to get past the French veto and let them know we won’t be competing for EU funds for space launch vehicles. Arianespace has already priced in the independent launch capability and it stands at 11-12 Ariane 6s a year. Anything less than that will result in a much higher price than the planned 40% cost reduction over Ariane 5. So when Arianespace comes short on that number what’s going to happen?

    Will the entire additional cost be covered by all EU members? Or will the EU content itself with launching BMW roadsters to Mars in attempt to use up those planned for but not paid for launch slots. I don’t underestimate the importance of space access for the EUs major military players, but without Britain there is only one of those, France. The question is will the rest of Europe chip in subsidize the program, or will they want the French to foot the bill? Or perhaps reorganize the entire European launch industry around a smaller and cheaper launcher such as an uprated Vega which would be good enough the vast majority of EU government satellites. So while I know space is important to the EU, its also true that everyone gets a vote and not everyone is going to want to kick in more money.

    Reply
  2. Right now Arianespace is struggling to fill their manifest. They are likely just trying to sell the remaining launch slots for their remaining production run of Ariane 5s, kind of like a car maker trying to get rid of the old model with fire sale prices before the new one is distributed. The real problem for them is not fire sale prices of Ariane 5 launch slots, but a lack of orders for the Ariane 6. Arianespace has let it be known that they need at least 4 more government purchases between 2020 and 2023 in order to issue the first production contract for 14 Ariane 6 vehicles.

    The entire Ariane 6 programs cost structure is premised around selling 11-12 launch vehicles a year and streamlining production enough to reduce the price by about 40 percent or the same discount they are offering for the last Ariane 5 slots. Its starting to look increasingly unlikely they will hit their goal and show that Ariane 6 can compete with Falcon 9 on cost let alone BFR/Starship or New Glenn. The fire sale pricing of Ariane 5 is therefore, just a bad omen of things to come. In the mid-2020s you will likely see the very limited commercial market dominated by reusable Falcon 9/Starship/New Glenn and every body else relying on government contracts. Arianespace, however is an odd bird in this crowd as it won’t have enough commercial customers to meet its goals and unlike Russia, China and the US, EU countries don’t have enough government missions to support Arianespace by themselves.

    Reply
  3. I thought they were wrapping up A5 production already and closing down the manufacturing line and supply chain? Are these newbuild, or more a forced GTO rideshare thing?

    Reply
  4. What is irrational is using the term ‘fair’ with regards to any adjective concerning the market.

    Market prices are what they are. There is nothing ‘fair’ about them. It is when people use the phrase ‘fair market [this or that]’ that socialists get their hook into pulling a contextual 180 that is actually the total antithesis of what markets do.

    The very word ‘fair’ is a subjective word, not an objective one.

    Reply
  5. Doesn’t make sense from a logistics perspective. You would need to move your maintenance and refurbishment facilities to Kourou which costs more money than launching from FL/TX and using a bit more fuel.

    TLDR: access to LEO is cheap enough that you don’t have to infinitely tune your launch architecture to put something in orbit.

    Reply
  6. What if we could combine the best launch site with the best rockets? Launch SpaceX rockets from Kourou & have two of SpaceX’s recovery barges in the Atlantic. One near due east for low inclination orbits & one near due north for polar orbits.
    This makes too much sense for the politics to work out.

    Reply
  7. I think you underestimate the importance of independent access to space for Europe. It is not stupidity. Same for Galileo satellites. And concerning subsidies in the USA before Space X, it has always been a known fact that the USA could develop its rocket industry thanks to a strong domestic government and military satellite market whereas Arianespace could never have access to this and so subsidized directly the cost in order to survive.

    Reply
  8. really? is that a serious question? .. exactly what subsidies are space x getting to compete for private sector launches from the US? you europeans and your corporate welfare socialism

    Reply
  9. More European subsidies to support their industry, just like their other foray into aerospace, Airbus. I doubt a US firm would launch with Ariane even with the cut because SpaceX will still be cheaper but if the Europeans are determined to subsidize industry that they feel is strategically important, it may be time for them to quiet down about steel tariffs etc that the US has implemented. Otherwise the US should target more European sectors with tariffs based on the state subsidy regimes they are running to compete in these strategic sectors.

    Reply
  10. The EU is determined to maintain an independent European launch capability. They’ll set the price so that they can get the launch cadence necessary to keep the program afloat. If that means that they have to subsidize their launchers a bit more, so be it.

    Reply

Leave a Comment