Liquid cooling: decisions, types, approaches.

By Nicole Hemsoth

June 11, 2012

Despite world economic turmoil of the recent severe European crisis, the demand for high performance computing services stays on the rise. Companies and institutions progressively see computational power as a source of competitive advantage and in many cases as the only optimal solution for many scientific and business challenges, from high energy physics to big data. This trend has brought an unprecedented rise in demand for high computational power. This is posing some sound energy and thermal management challenges.

The energy problem in data centres is two sided. On the one hand, data centres have a problem of energy consumption, which enlarges bills also in countries where the cost of energy is relatively low. On the other hand, there is a problem of peak power demand, so, in other words, a problem of availability.  Megawatt installations are not so uncommon anymore, meaning that a request of power similar to the one that has traditionally belonged to the heavy industry sector is becoming almost the norm, in some occasions requiring special arrangement for power systems and electrical lines.

Thermal management is exacerbated by another trend: density. In many cases, rack powers of 30 kW are well beyond what legacy air cooling can handle. In the modern HPC, the high powers in play often leave few options but resorting to some form of water cooling.

Liquid cooling has many advantages, which derive from the much higher heat capacity per unit volume of water compared to air (we are talking about a factor of 3500 times higher). Liquid cooling implies higher densities, energy savings and the possibility to reuse the thermal energy that the water extracts from the IT equipment. Some additional advantages can be found in terms of lower noise levels, less vibrations and close control of electronics temperatures.

The best approach in deciding what type of cooling to implement is to consider alternatives in relation to technical and business needs, the type of air and liquid cooling system available within budget and a series of variables that play an important role in the decision: the desired density versus space availability, new construction versus existing construction, the proximity to natural sources of cold water like rivers and lakes, the local climate, the cost of energy and the thermal energy recovery possibilities.

For instance, high performance high density requirements may leave little choice than liquid cooling to efficiently manage the extraction of the heat from the supercomputers. While, if the data centre has an economizer and the climate is best suited to air-side economizers (mild temperatures and moderate humidity) than an air cooled DC may have more sense. 

Deciding the cooling system may also take in consideration the type of water cooling to be installed. There are solutions that simply create an extension of the existing liquid-cooling loop closer to the IT equipment like in the case of liquid cooled racks (liquid cooled door, closed-liquid rack). In other solutions, in-row units are embedded in rows of data center cabinets, providing localized air distribution and management. Alternatively, overhead cooling suspends from the ceiling complements a hot aisle/cold aisle arrangement. As hot air rises from the hot aisle, the overhead cooler captures it, conditions it, and releases it back to the cold aisle

More effective cooling can be reached when the liquid is brought in the near proximity of the electronic components like in the case of submerged cooling, spray cooling or direct (embedded) cooling.

In the first case, the electronic components are immerged in oil and water which is kept in circulation through small pumps. In the second, the water is vaporized and tiny drops of water fall on the electronics evaporating immediately and taking away a lot of heat. In the latter, water is taken through metal plates or micro pipes to direct contact with processors, memory and other components.

Another distinction is normally made between hot and cold liquid cooling. The definition of hot liquid cooling can be vary. In Eurotech we think that hot liquid cooling means the technology capable of using a liquid (e.g. water) with a temperature above the server room temperature.  We also accept that, pushing the bar up in terms of max coolant temperature, hot liquid cooling may take place when the water is hot enough to allow thermal energy reuse.

In any kind of liquid cooling, one aspect that needs careful attention is the risk of leaking. This is an issue because the electronic components are upgraded on a routine basis resulting in many systems with the need to disconnect and reconnect the liquid carrying lines. Also, there is the need to consider whether cooling with water brings on all of its potential. For instance, resorting to chillers to cool the water will allow density, but limit the energy savings that are maximized with hot water cooling technologies, thanks to air conditioning avoidance. However, it is no news that new powerful processors with TDP of 150W may require coolant temperatures lower that the ones guaranteed by free cooling in warm climates. An additional downside of increasing water temperature may be the higher operating temperature of electronic components.  This risk needs to be balanced by the advantages coming from levelling temperatures on the mother board and avoiding hot spots at data center level.

Eurotech approach

Eurotech has developed liquid cooling systems for more than 7 years and it was the first in the market to offer a hot liquid cooling with high serviceability. Eurotech Aurora supercomputers have been liquid cooled since product one and day one, allowing for precious competences and know how to be waived within the fabric of the organization. This experience helped the development of our idea of liquid cooling.

Eurotech liquid cooling is:

Hot. That means using hot water of 50+ °C, balancing customer needs, density targets, data center temperature and site temperature/humidity profiles. Eurotech delivers to customers the liquid cooling solution that allows utilizing the water at the maximum temperature possible across the year. 

Direct.  The cooling takes place inside the rack, where aluminum cold plates are put in direct contact with the components, allowing to maximize the heat transfer and heat extraction efficacy. The good side effect is to level out temperatures on board avoiding hot spots.

Green. Eurotech aims to utilize free coolers (liquid to air heat exchangers) in any climate zone. Solutions are designed to avoid air conditioning, while maintaining the highest density possible, and to exploit, if required and wherever it is possible, thermal energy recovery.

Comprehensive. The “cold plates” cool processors, memory, FPGAs, power supply, switches and any other heat generating component, including GPUs or other accelerators. This means that there is not a single heat source in the rack that is not cooled, preventing hot spots at DC level.

Serviceable. Eurotech Aurora HPC boards are hot swappable despite being water cooled thanks to connectors that seal instantaneously when a node card is extracted for maintenance or management purposes. The node cards are blades that a single person can easily manage.

Safe. Eurotech understand that it is imperative to keep water away from electronics. For this reason we have spent several years to develop a system that doesn’t leak and to mature those competencies that guide our customers into the correct and trouble free maintenance of the liquid cooling infrastructure.

Indeed, one of the Eurotech focus is on correct liquid cooling operations and maintenance, which is fundamental to preserve the system safety and integrity and keep performances at top levels.

“The maintenance of liquid cooling systems is not a daunting task” says Paul Arts, Eurotech technical director “but it requires following guidelines many of them are conveniently collected by Ashrae. At Eurotech, we assist our customers in approaching hot water cooling, designing the systems and training the customers in operations and maintenance. If have to spare my 2 cents, areas I would focus my attention are water quality, anti-corrosion precautions, flow rate and dew point temperatures”

Eurotech has experienced that correct operations maximize the life not only of the cooling system but also of the electronic components, rounding up the advantages of using hot water cooling. Eurotech believes in liquid cooling as an approachable and concrete solution for facing energy and thermal issues, especially in those contexts that are climatically unfavourable.

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!

Quantum Watchers – Terrific Interview with Caltech’s John Preskill by CERN

July 17, 2024

In case you missed it, there's a fascinating interview with John Preskill, the prominent Caltech physicist and pioneering quantum computing researcher that was recently posted by CERN’s department of experimental physi Read more…

Aurora AI-Driven Atmosphere Model is 5,000x Faster Than Traditional Systems

July 16, 2024

While the onset of human-driven climate change brings with it many horrors, the increase in the frequency and strength of storms poses an enormous threat to communities across the globe. As climate change is warming ocea Read more…

Researchers Say Memory Bandwidth and NVLink Speeds in Hopper Not So Simple

July 15, 2024

Researchers measured the real-world bandwidth of Nvidia's Grace Hopper superchip, with the chip-to-chip interconnect results falling well short of theoretical claims. A paper published on July 10 by researchers in the U. Read more…

Belt-Tightening in Store for Most Federal FY25 Science Budets

July 15, 2024

If it’s summer, it’s federal budgeting time, not to mention an election year as well. There’s an excellent summary of the curent state of FY25 efforts reported in AIP’s policy FYI: Science Policy News. Belt-tight Read more…

Peter Shor Wins IEEE 2025 Shannon Award

July 15, 2024

Peter Shor, the MIT mathematician whose ‘Shor’s algorithm’ sent shivers of fear through the encryption community and helped galvanize ongoing efforts to build quantum computers, has been named the 2025 winner of th Read more…

Weekly Wire Roundup: July 8-July 12, 2024

July 12, 2024

HPC news can get pretty sleepy in June and July, but this week saw a bump in activity midweek as Americans realized they still had work to do after the previous holiday weekend. The world outside the United States also s Read more…

Aurora AI-Driven Atmosphere Model is 5,000x Faster Than Traditional Systems

July 16, 2024

While the onset of human-driven climate change brings with it many horrors, the increase in the frequency and strength of storms poses an enormous threat to com Read more…

Shutterstock 1886124835

Researchers Say Memory Bandwidth and NVLink Speeds in Hopper Not So Simple

July 15, 2024

Researchers measured the real-world bandwidth of Nvidia's Grace Hopper superchip, with the chip-to-chip interconnect results falling well short of theoretical c Read more…

Shutterstock 2203611339

NSF Issues Next Solicitation and More Detail on National Quantum Virtual Laboratory

July 10, 2024

After percolating for roughly a year, NSF has issued the next solicitation for the National Quantum Virtual Lab program — this one focused on design and imple Read more…

NCSA’s SEAS Team Keeps APACE of AlphaFold2

July 9, 2024

High-performance computing (HPC) can often be challenging for researchers to use because it requires expertise in working with large datasets, scaling the softw Read more…

Anders Jensen on Europe’s Plan for AI-optimized Supercomputers, Welcoming the UK, and More

July 8, 2024

The recent ISC24 conference in Hamburg showcased LUMI and other leadership-class supercomputers co-funded by the EuroHPC Joint Undertaking (JU), including three Read more…

Generative AI to Account for 1.5% of World’s Power Consumption by 2029

July 8, 2024

Generative AI will take on a larger chunk of the world's power consumption to keep up with the hefty hardware requirements to run applications. "AI chips repres Read more…

US Senators Propose $32 Billion in Annual AI Spending, but Critics Remain Unconvinced

July 5, 2024

Senate leader, Chuck Schumer, and three colleagues want the US government to spend at least $32 billion annually by 2026 for non-defense related AI systems.  T Read more…

Point and Click HPC: High-Performance Desktops

July 3, 2024

Recently, an interesting paper appeared on Arvix called Use Cases for High-Performance Research Desktops. To be clear, the term desktop in this context does not Read more…

Atos Outlines Plans to Get Acquired, and a Path Forward

May 21, 2024

Atos – via its subsidiary Eviden – is the second major supercomputer maker outside of HPE, while others have largely dropped out. The lack of integrators and Atos' financial turmoil have the HPC market worried. If Atos goes under, HPE will be the only major option for building large-scale systems. Read more…

Everyone Except Nvidia Forms Ultra Accelerator Link (UALink) Consortium

May 30, 2024

Consider the GPU. An island of SIMD greatness that makes light work of matrix math. Originally designed to rapidly paint dots on a computer monitor, it was then Read more…

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…


Nvidia Economics: Make $5-$7 for Every $1 Spent on GPUs

June 30, 2024

Nvidia is saying that companies could make $5 to $7 for every $1 invested in GPUs over a four-year period. Customers are investing billions in new Nvidia hardwa Read more…

Nvidia Shipped 3.76 Million Data-center GPUs in 2023, According to Study

June 10, 2024

Nvidia had an explosive 2023 in data-center GPU shipments, which totaled roughly 3.76 million units, according to a study conducted by semiconductor analyst fir Read more…

AMD Clears Up Messy GPU Roadmap, Upgrades Chips Annually

June 3, 2024

In the world of AI, there's a desperate search for an alternative to Nvidia's GPUs, and AMD is stepping up to the plate. AMD detailed its updated GPU roadmap, w Read more…

Some Reasons Why Aurora Didn’t Take First Place in the Top500 List

May 15, 2024

The makers of the Aurora supercomputer, which is housed at the Argonne National Laboratory, gave some reasons why the system didn't make the top spot on the Top Read more…

Intel’s Next-gen Falcon Shores Coming Out in Late 2025 

April 30, 2024

It's a long wait for customers hanging on for Intel's next-generation GPU, Falcon Shores, which will be released in late 2025.  "Then we have a rich, a very Read more…

Leading Solution Providers


Google Announces Sixth-generation AI Chip, a TPU Called Trillium

May 17, 2024

On Tuesday May 14th, Google announced its sixth-generation TPU (tensor processing unit) called Trillium.  The chip, essentially a TPU v6, is the company's l 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…

IonQ Plots Path to Commercial (Quantum) Advantage

July 2, 2024

IonQ, the trapped ion quantum computing specialist, delivered a progress report last week firming up 2024/25 product goals and reviewing its technology roadmap. 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…

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…

The NASA Black Hole Plunge

May 7, 2024

We have all thought about it. No one has done it, but now, thanks to HPC, we see what it looks like. Hold on to your feet because NASA has released videos of wh 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…

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…

  • arrow
  • Click Here for More Headlines
  • arrow