Startup Makes Liquid Cooling an Immersive Experience

By Michael Feldman

August 31, 2010

There’s nothing like a blazing hot summer to focus one’s attention on the best ways to keep cool. That goes for datacenter operators as well, who are equally worried about keeping their servers properly chilled. While there is no shortage of innovative cooling solutions being proffered by various vendors, a new liquid immersion cooling solution from startup Green Revolution Cooling could end up being the best of them all.

The stakes for more efficient datacenter cooling are already high. Power consumption for a traditional air-cooled facility eats up a third to more than a half of the energy cost. Making cooling more efficient leaves more money available for computing, which, after all, is the central purpose of the datacenter. Efficient cooling is an especially important consideration in high performance computing, since this class of users gravitate toward faster and denser (and thus hotter) server configurations. If the setup in the center is not optimal, you end up sacrificing a lot of FLOPS for cooling.

With the increasing density of servers, storage, switches and other equipment, facility managers are taking an extra hard look at liquid cooling. Water-cooled servers have been around for decades, and direct-cooled CPUs are now being offered by a handful of vendors. Submerged liquid cooling, too, has been around since the days of the Cray 2, but this technology may be poised for a big comeback.

Servers Take a Bath

Green Revolution Cooling (GRC), a two year-old company based in Austin, Texas, is offering a general-purpose liquid immersion cooling solution that they introduced at SC09 in Portland last November. It was selected as one of the “Disruptive Technologies of the Year” for the 2009 conference, an award they’ve recaptured for SC10.

In a nutshell, the system consists of a 42U rack enclosure tipped on its back and filled with an inert mineral oil mixture in which you immerse the server hardware. A pump is used to circulate the oil to an external heat exchanger, typically located outside the building.

The big advantage is that, unlike water, the oil formulation is not electrically conductive, but has 1,200 times the heat capacity of air. And since the oil is in direct contact with all the components, it only needs to be cooled down to about 104F (40C) to be effective. (CPUs can operate at 75C and hard drives at 45C.) Unless your datacenter happens to be located in Yuma, Arizona, cooling a liquid to 40C is relatively easy to attain with a simple heat exchanger or cooling tower. The solution is advertised to reduce the cooling energy by 90 percent and cut overall power consumption in the datacenter by up to 45 percent. The pitch is that a single 10kW server rack at 8 cents per kWh will save over $5,000 per year on energy costs alone.

According to Green Revolution co-founder Christiaan Best, basically any piece of datacenter equipment — rackmount server, blade, switch — that adheres to the standard 19-inch form factor can be slid into the GRC enclosure. The only equipment modifications required are the removal of the internal fans (you don’t need air cooling any more) and the sealing of any hard drive units, with an epoxy coating, to make them airtight. Typically this procedure takes a few minutes per server.

Because the GRC enclosure is laid on its back, it does takes up more floor space than a regular vertical rack. But since you no longer need hot aisles, chillers, and CRAC units, there is extra square footage to play with. Also, because there is no need to run cold air beneath the equipment anymore, the raised floor is now superfluous. “Essentially you could run it in a barn,” says Best. “All you need is a level floor.”

If you’re looking for performance, the GRC rack allows you to overclock the processors without worrying about melting the server. An NSF-funded study found that cranking up the clock on an Intel E5520 “Nehalem” CPU inside a GRC-cooled server yielded a 54 percent performance boost on Linpack, while keeping the CPU temperature at 76C. The server cost per gigaflop was reduced by about 50 percent.

It’s not just for overclocking. Theoretically, you could throw almost any sort of artificially dense board — multi-GPUs servers, custom blades with 10 CPUs on the motherboard, etc. — into the oil bath and realize the additional cost benefit of shrinking down your hardware footprint.

One possible roadblock to widespread adoption is the lack of warranty support from the OEMs. Warranties don’t typically allow the customer to take the server apart and dunk it into foreign liquids. According to Best, they’ve been talking with all the major OEMs to get their solution qualified under the original warranties, but currently none have committed to supporting the GRC setup. Since many of the big system vendors have their own liquid cooling solutions they’d like sell, they are likely to be less than enthusiastic to qualify a third-party solution.

In any case, Best says they’ve retained third-party support that will honor the original equipment warranties, so customers can be covered for any mishaps. GRC has logged over a quarter million server hours on their in-house test system and has yet to encounter a failure (with the exception of hard drive mechanical failures). Although there is no data to support it, Best is fairly certain that their solution will extend the life of the servers, given the more stable thermal environment, the lack of vibration from internal fans, and the elimination of oxidation on the electrical contacts.

Looking for a Few Brave Customers

Austin-based Midas Networks, a collocation firm, is the company’s first customer. Midas has purchased four of the GRC racks, and the systems are scheduled to be up and running later this year. Best says they also have a number of other customers in the pipeline, including some with HPC facilities, but no checks are in the bank just yet.

With the exception of Green Revolution itself, the Texas Advanced Computing Center (TACC) has acquired the most experience with the technology. TACC installed a pre-production GRC unit back in April and has been putting the system through its paces for the past five months.

Even in oil-rich Texas, energy is not cheap, so power savings has become a big priority at TACC. “We’re really, really chill-water limited where we are now,” says Dan Stanzione, TACC’s deputy director. According to him, they don’t have the ability to add any more chilled water capacity, but do have plans to expand computing capability over the next several years.

The TACC experiment started with immersing some older 1U servers in the GRC enclosure, and since then they’ve added other equipment including InfiniBand switches, GPU-powered servers, and blades. According to Stanzione, all the hardware has performed flawlessly, with no failures to date. They’ve even overclocked some of the server CPUs by 30 to 40 percent, without incident.

At present they have about 10kW of equipment in the rack, and are using just 250 watts to power the GRC solution. That’s more than a 90 percent reduction when compared to the 3,000 to 4,000 watts they would have consumed with a conventional air-cooled system. Stanzione estimates the total power savings for the whole system (equipment plus cooling) was reduced by 25 to 30 percent. “The overall power consumption has been fantastic,” he says.

The TACC crew is going to continue collecting data with the GRC system for the rest of the year. If everything checks out, Stanzione would like to start putting some production units into the upcoming datacenter buildout. They’re already thinking about loading 30 to 40 kW of compute equipment into a single rack, and GRC cooling would make that level of density quite practical. Further into the future, Stanzione is thinking about the cost savings they could accrue by immersing all 140 racks of the center’s equipment. “I think this has a tremendous amount of potential,” he says.

Barring some unforseen technological breakthrough, datacenter computing is only going to get denser and hotter in the years ahead. And since the cooling capacity of air isn’t going to change, the move to liquid-cooled systems appears all but inevitable. “You may not buy liquid cooling from us,” concludes Best, “but you will buy it from someone.”

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion XL — were added to the benchmark suite as MLPerf continues Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa Read more…

MLPerf Inference 4.0 Results Showcase GenAI; Nvidia Still Dominates

March 28, 2024

There were no startling surprises in the latest MLPerf Inference benchmark (4.0) results released yesterday. Two new workloads — Llama 2 and Stable Diffusion Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Shutterstock 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire