Quantum Computer Hosting – Rumours from the Trade Show Floor

HPC GPU


Psst, please only share this enterprise quantum computing insight with your HPC enthusiast friends. My boss would have a bird if he thought I was leaning 2+ years into the future with enterprise quantum computer hosting material!

During our recent HPC & AI Field Trip to Iceland it became increasingly clear that a compelling competitive advantage can be garnered by cascading best-in-class compute technologies to solve formerly impossible tasks. Conventional CPUs with their flexibility will control the tasks and feed large amounts of number crunching to GPUs, FPGAs or other accelerators, but an additional technology is needed for tasks too large for conventional deterministic logic – quantum computers.

Today quantum computers are mostly in the research labs of their designers, sometimes with the ability for outsiders to test their own quantum logic, beta-test like. Over the next few years these devices may harden sufficiently for individual companies to acquire and operate them - likely adjacent to their existing industrial scale HPC clusters, many of which we host here in Iceland. There are at least 3 types of quantum computer: semi-conductor, electromagnetic, and atomic scale. The superconducting semi-conductor ones currently appear to be the most advanced. My curiosity got the better of me and I recently visited Janis, here in Boston, to learn all about the environmentals necessary for semi-conductor-based quantum computing requiring 7mK (7/1000⁰C above absolute zero temperature, - 272.993⁰C). Vladimir and Ashley were my hosts for the afternoon, and they were gifted at explaining bleeding-edge cryogenics technology to a rusty EE major. It was a seriously fascinating afternoon out, and here are some of my insights.

Modern cryogenics equipment reminds me of my English great grandfather’s old country house: heat in the kitchen from an Aga wood burning stove and every adjoining room subsequently cooler until my play room which was frigid in winter. Cryogenic equipment is generally orientated vertically with the cold spot at the bottom because cold gas sinks and the pumps/coolers work best in this plane. Like my great grandfather’s house, the coolers have layers of decreasing temperature, some of the layers are to limit the reach of ambient temperature, and some are part of the cooling process.

The 7 layers of the Janis Quantum Computer Cryostat – Photo Janis Research


Unlike my great grandfather’s house, the layers of the cryostat are tightly sealed and designed to stay that way over a wide temperature range. The whole unit fits in a large very-well insulated can with the cold ~48cm-wide and 25cm-deep cold chamber below the bottom plate. The heat plates are gold-plated very pure copper with holes for the instrumentation wires to transition to the equipment. The areas around the wires are tightly sealed, and any unused hole has a blanking plate fitted. Infrared energy and magnetism are the enemies of cryostats, so there is no ferrous material or any direct paths for electromagnetic energy to pass between the cryostat layers. Interestingly, the 6Ghz signal wires going down to the quantum computer are hybrid alloy at the top and supercomputing at the bottom. With up to 500 possible and a cost of $2,000 each they match the cost of the cooling infrastructure!

The cryostat has two cooling engines, an external two-stage pulse tube refrigerator which can generate temperatures down to about 2K, and an internal 3He/4He dilution refrigerator generating theoretical temperatures down to about 2mK if the mixing chambers have sufficient thermal conductivity. This quantum pulse tube is configured for 3K and the dilution refrigerator for 7mK, a major achievement for a production product.

There are two operational modes for a cryostat: pre-cooling and cooling. The pre-cooling mode starts at room temperature with liquid nitrogen flowing into the blue area below, and the heat switches in the red area between the layers being connected. Over a day or so the whole temperature quickly reduces to 77K. Thereafter, the pulse tube refrigerator starts cooling the third layer down to 3K. Once the cryostat third layer stabilises at 3K, the layers above it stabilise at 10K and 30K to limit the influence of the outside ambient temperature.

Liquid nitrogen and heat switches to accelerate the precooling phase – photo Janis Research


Once the lower cryostat is stabilized at 3K, the 3He/4He dilution refrigerator starts to cool these lower layers and the heat switches disconnect the lower layers, which takes about another 12 hours to reach 7mK throughout the quantum computer space. The whole process is automated and dry, meaning that there is little or no cryogenic gas consumption.

A fully operational cryostat with its control electronics and the pulse tube cooler in the right – photo Janis Research


Operating a cryostat in a computer facility requires about 25kW of power and connectivity to cooling water for the pulse tube refrigerator. It will also benefit from a low vibration solid floor and accessibility to its HPC compute cluster. Verne Global can, when they are available, closely link your future quantum computer to your hosted compute infrastructure or, alternatively, its equivalent running in our bare metal cloud.

During my research I discovered a multitude of quantum computer trivia, the most interesting of which are:

  • The cryogenic heat switches are based on two materials with large differences in the coefficients of thermal expansion configured for a switching temperature – like a bi-metallic thermostat
  • A full complement of quantum computer interconnect wiring will cost as much, if not more than, the cooling plant
  • A 20L charge of 3He will cost about $40,000, but the 100L of 4He will only cost about $2,000
  • 3He is mostly recovered during the servicing of nuclear warheads
  • The dilution refrigerator mixer surface area ratio needed for a factor of 2 lower dilution refrigerator temperatures increases by a factor of 214(16,384)
  • Quantum computer jobs usually run for about 200µS before their noise errors become excessive – the cooling plant is not a factor in this
  • Ex-semi-conductor fab technicians are ideal white gloves data center staff due to their ability to diagnose/fix gas leaks

Get a head of the quantum computing curve and host your HPC GPU clusters, simulating quantum compute algorithms, with Verne Global today.


Written by Bob Fletcher

See Bob Fletcher's blog

Bob, a veteran of the telecommunications and technology industries, is Verne Global's VP of Strategy. He has a keen interest in HPC and the continuing evolution of AI and deep neural networks (DNN). He's based in Boston, Massachusetts.

Related blogs

Qubits North America - Rumours from the Trade Show Floor

It's lunchtime here in the beautiful city of Amsterdam where I am enjoying my time at the excellent World Summit AI, at which Verne Global are exhibiting, and earlier today our CTO Tate Cantrell moderated an engaging panel on today's hottest AI disruptors. As I take a breather from the networking, I thought I would write some thoughts about another excellent event which I went to a couple of weeks ago - Qubits North America.

Read more


CERN - When science can't use the cloud

Demand for high performance computing (HPC) is growing fast - and you might expect it to become kind of generic. But leading research sites like CERN still make extreme demands.

Read more


DevOps. Stacking up to be a Common Theme at ISC18

Yesterday was a good day to be in Frankfurt. All of the majors in the supercomputing universe descended upon the Messe Frankfurt to begin ISC18 with a series of training seminars. For my morning session, I chose Getting Started with Containers on HPC through Singularity put on by the team at Sylabs, Inc. I have been tracking the progress on Singularity in the HPC community since before Sylabs was founded by CEO Greg Kurtzer in an effort to bring the technology of root secure containers into the realm of enterprise supported software. I was excited to hear about the progress that Sylabs has made and to see where the future of containers lies for the broader HPC community. If I was forced to sum the tutorial into a single portmanteau, it would be DevOps. After this session, it is clear to me that the world of DevOps that has been created in the cloud native universe is on a collision course with HPC. And the future of science says that it can’t happen soon enough.

Read more

We use cookies to ensure we give you the best experience on our website, to analyse our website traffic, and to understand where our visitors are coming from. By browsing our website, you consent to our use of cookies and other tracking technologies. Read our Privacy Policy for more information.