Key Considerations in the Cooling of AI and HPC Systems

November 19, 2018

The critical driver for any AI or HPC cluster is the ability to fit to a usage model that requires continuous computing throughput across the entire system.   HPC and the latest AI architectures are characterized by clusters and their nodes running at 100% utilization for sustained periods. Further, as any such application is always compute limited, cutting edge AI and HPC requires the highest performance versions of the latest CPUs and GPUs. This means the highest frequency offerings of processors such as NVIDIA’s V100 GPUs and Intel’s Xeon Scalable Processors (Skylake).   Indeed, the highest performance versions also means the highest wattages.

The wattages for NVIDIA’s Volta V100 GPU are currently at 300 watts and both Intel’s Xeon Scalable Processors (Skylake) CPU and Xeon Phi (KNM) MIC-styled GPU & have been publicly announced at 205 and 320 watts, respectively.

These chip wattages translate into substantially higher wattage densities at both the node-level and rack-level not simply because of the component wattages alone but also due to the requirement for the shortest possible signal distance between processors, GPUs, switches both in and between cluster racks.  For HPC workloads, overall computing throughput is also constrained by signal delay both in the spatial and time domains.  These factors are driving cluster racks well beyond 50kW to 80kW or higher.

Racks with air heat sinks struggle to handle the heat to maintain this maximum throughput and CPU throttling occurs due to inefficient air cooling. Particularly for HPC sites, reducing rack densities can mean increasing interconnect distances resulting in greater latency and lower cluster throughput.  As a result, liquid cooling is required for the highest performance AI and HPC systems.

Implementation of liquid cooling at its best requires an architecture that is highly reliable, flexible to a variety of heat rejection scenarios, adaptable to the underlying compute and interconnect architecture and is not cost prohibitive.  All of these factors must be addressed by the liquid cooling architecture so that compromises need not be made in the compute architecture.

From a reliability standpoint, there are a number of cooling elements to consider.  The most important factor is to have very low pressures throughout the system (e.g. 4psi) both in the node and at the rack level which also addresses the cost issue in not requiring high-pressure components and allows for higher densities. Additionally, redundancy in the cooling architecture is needed to mitigate any impact of single point failure and to allow for component replacement during normal scheduled down-time.  Finally, a robust monitoring and alarming system is needed to monitor the cooling system and anticipate potential issues over time.

Adaptability of the cooling architecture is often overlooked largely due to the historical approach of liquid cooling related to the use of high-pressure pumping systems that take a one-size-fits-all approach.  With Asetek’s distributed architecture in contrast, coolers (integrated pumps/cold plates) are placed within server or blade nodes, and these coolers replace CPU/GPU air heat sinks in order to remove heat with warm water.   This has numerous advantages.

Heat capture with distributed pumping liquid cooling combined with options for heat rejection into either data centre air or facilities liquid seen in Asetek’s architecture provides a flexibility to address high-wattage on an as-needed basis. The Asetek architecture heat rejection options provide adaption to existing air-cooled data centers and to liquid-cooled facilities.

Adding liquid cooling with no impact on data center infrastructure can be done with Asetek’s InRackLAAC™, a server-level Liquid Assisted Air Cooling (LAAC) option.   With InRackLAAC the redundant liquid pump/cold plates are paired with a shared HEX (radiator) in the rack.   Via the HEX the captured heat is exhausted into the data center. InRackLAAC places a shared HEX with a 6kW 2U chassis that is connected to a “block” of up to 12 servers.   Existing data center HVAC systems handle the heat.

Multiple such computing blocks can be used in a rack. InRackLACC allows incorporation of the highest wattage CPUs/GPUs and additionally racks can contain a mix of liquid-cooled and air-cooled nodes.

When facilities water is routed to the racks, Asetek’s 80kW InRackCDU™ D2C (Direct-to-Chip) can capture 60 to 80 percent of server heat into liquid, reducing data center cooling costs by over 50 percent and allowing 2.5x-5x increases in data center server density.  Because hot water (up to 40ºC) is used to cool, it does not require expensive HVAC systems and can utilize inexpensive dry coolers

With InRackCDU the heat collected is moved via a sealed liquid path to heat exchangers for transfer of heat into facilities water.  InRackCDU is mounted in the rack along with servers.  Using 4U, it connects to nodes via Zero-U PDU style manifolds in the rack.

Asetek’s distributed pumping architecture at the server, rack, cluster and site levels delivers flexibility in the areas of heat capture, coolant distribution and heat rejection that other approaches do not.

Visit Asetek.com to learn more.

 

Subscribe to HPCwire's Weekly Update!

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

AWS Introduces a Flurry of New EC2 Instances at re:Invent

November 30, 2022

AWS has announced three new Amazon Elastic Compute Cloud (Amazon EC2) instances powered by AWS-designed chips, as well as several new Intel-powered instances – including ones targeting HPC – at its AWS re:Invent 2022 Read more…

Quantum Riches and Hardware Diversity Are Discouraging Collaboration

November 28, 2022

Quantum computing is viewed as a technology for generations, and the spoils for the winners are huge, but the diversity of technology is discouraging collaboration, an Intel executive said last week. There are close t Read more…

2022 Road Trip: NASA Ames Takes Off

November 25, 2022

I left Dallas very early Friday morning after the conclusion of SC22. I had a race with the devil to get from Dallas to Mountain View, Calif., by Sunday. According to Google Maps, this 1,957 mile jaunt would be the longe Read more…

2022 Road Trip: Sandia Brain Trust Sounds Off

November 24, 2022

As the 2022 Great American Supercomputing Road Trip carries on, it’s Sandia’s turn. It was a bright sunny day when I rolled into Albuquerque after a high-speed run from Los Alamos National Laboratory. My interview su Read more…

2022 HPC Road Trip: Los Alamos

November 23, 2022

With SC22 in the rearview mirror, it’s time to get back to the 2022 Great American Supercomputing Road Trip. To refresh everyone’s memory, I jumped in the car on November 3rd and headed towards SC22 in Dallas, stoppi Read more…

AWS Solution Channel

Shutterstock 110419589

Thank you for visiting AWS at SC22

Accelerate high performance computing (HPC) solutions with AWS. We make extreme-scale compute possible so that you can solve some of the world’s toughest environmental, social, health, and scientific challenges. Read more…

 

shutterstock_1431394361

AI and the need for purpose-built cloud infrastructure

Modern AI solutions augment human understanding, preferences, intent, and even spoken language. AI improves our knowledge and understanding by delivering faster, more informed insights that fuel transformation beyond anything previously imagined. Read more…

Chipmakers Looking at New Architecture to Drive Computing Ahead

November 23, 2022

The ability to scale current computing designs is reaching a breaking point, and chipmakers such as Intel, Qualcomm and AMD are putting their brains together on an alternate architecture to push computing forward. The chipmakers are coalescing around the new concept of sparse computing, which involves bringing computing to data... Read more…

AWS Introduces a Flurry of New EC2 Instances at re:Invent

November 30, 2022

AWS has announced three new Amazon Elastic Compute Cloud (Amazon EC2) instances powered by AWS-designed chips, as well as several new Intel-powered instances Read more…

Quantum Riches and Hardware Diversity Are Discouraging Collaboration

November 28, 2022

Quantum computing is viewed as a technology for generations, and the spoils for the winners are huge, but the diversity of technology is discouraging collaborat Read more…

2022 HPC Road Trip: Los Alamos

November 23, 2022

With SC22 in the rearview mirror, it’s time to get back to the 2022 Great American Supercomputing Road Trip. To refresh everyone’s memory, I jumped in the c Read more…

QuEra’s Quest: Build a Flexible Neutral Atom-based Quantum Computer

November 23, 2022

Last month, QuEra Computing began providing access to its 256-qubit, neutral atom-based quantum system, Aquila, from Amazon Braket. Founded in 2018, and built o Read more…

SC22’s ‘HPC Accelerates’ Plenary Stresses Need for Collaboration

November 21, 2022

Every year, SC has a theme. For SC22 – held last week in Dallas – it was “HPC Accelerates”: a theme that conference chair Candace Culhane said reflected Read more…

Quantum – Are We There (or Close) Yet? No, Says the Panel

November 19, 2022

For all of its politeness, a fascinating panel on the last day of SC22 – Quantum Computing: A Future for HPC Acceleration? – mostly served to illustrate the Read more…

RISC-V Is Far from Being an Alternative to x86 and Arm in HPC

November 18, 2022

One of the original RISC-V designers this week boldly predicted that the open architecture will surpass rival chip architectures in performance. "The prediction is two or three years we'll be surpassing your architectures and available performance with... Read more…

Gordon Bell Special Prize Goes to LLM-Based Covid Variant Prediction

November 17, 2022

For three years running, ACM has awarded not only its long-standing Gordon Bell Prize (read more about this year’s winner here!) but also its Gordon Bell Spec Read more…

Nvidia Shuts Out RISC-V Software Support for GPUs 

September 23, 2022

Nvidia is not interested in bringing software support to its GPUs for the RISC-V architecture despite being an early adopter of the open-source technology in its GPU controllers. Nvidia has no plans to add RISC-V support for CUDA, which is the proprietary GPU software platform, a company representative... Read more…

RISC-V Is Far from Being an Alternative to x86 and Arm in HPC

November 18, 2022

One of the original RISC-V designers this week boldly predicted that the open architecture will surpass rival chip architectures in performance. "The prediction is two or three years we'll be surpassing your architectures and available performance with... Read more…

AWS Takes the Short and Long View of Quantum Computing

August 30, 2022

It is perhaps not surprising that the big cloud providers – a poor term really – have jumped into quantum computing. Amazon, Microsoft Azure, Google, and th Read more…

Chinese Startup Biren Details BR100 GPU

August 22, 2022

Amid the high-performance GPU turf tussle between AMD and Nvidia (and soon, Intel), a new, China-based player is emerging: Biren Technology, founded in 2019 and headquartered in Shanghai. At Hot Chips 34, Biren co-founder and president Lingjie Xu and Biren CTO Mike Hong took the (virtual) stage to detail the company’s inaugural product: the Biren BR100 general-purpose GPU (GPGPU). “It is my honor to present... Read more…

AMD Thrives in Servers amid Intel Restructuring, Layoffs

November 12, 2022

Chipmakers regularly indulge in a game of brinkmanship, with an example being Intel and AMD trying to upstage one another with server chip launches this week. But each of those companies are in different positions, with AMD playing its traditional role of a scrappy underdog trying to unseat the behemoth Intel... Read more…

Tesla Bulks Up Its GPU-Powered AI Super – Is Dojo Next?

August 16, 2022

Tesla has revealed that its biggest in-house AI supercomputer – which we wrote about last year – now has a total of 7,360 A100 GPUs, a nearly 28 percent uplift from its previous total of 5,760 GPUs. That’s enough GPU oomph for a top seven spot on the Top500, although the tech company best known for its electric vehicles has not publicly benchmarked the system. If it had, it would... Read more…

JPMorgan Chase Bets Big on Quantum Computing

October 12, 2022

Most talk about quantum computing today, at least in HPC circles, focuses on advancing technology and the hurdles that remain. There are plenty of the latter. F Read more…

Using Exascale Supercomputers to Make Clean Fusion Energy Possible

September 2, 2022

Fusion, the nuclear reaction that powers the Sun and the stars, has incredible potential as a source of safe, carbon-free and essentially limitless energy. But Read more…

Leading Solution Providers

Contributors

UCIe Consortium Incorporates, Nvidia and Alibaba Round Out Board

August 2, 2022

The Universal Chiplet Interconnect Express (UCIe) consortium is moving ahead with its effort to standardize a universal interconnect at the package level. The c Read more…

Nvidia, Qualcomm Shine in MLPerf Inference; Intel’s Sapphire Rapids Makes an Appearance.

September 8, 2022

The steady maturation of MLCommons/MLPerf as an AI benchmarking tool was apparent in today’s release of MLPerf v2.1 Inference results. Twenty-one organization Read more…

SC22 Unveils ACM Gordon Bell Prize Finalists

August 12, 2022

Courtesy of the schedule for the SC22 conference, we now have our first glimpse at the finalists for this year’s coveted Gordon Bell Prize. The Gordon Bell Pr Read more…

Not Just Cash for Chips – The New Chips and Science Act Boosts NSF, DOE, NIST

August 3, 2022

After two-plus years of contentious debate, several different names, and final passage by the House (243-187) and Senate (64-33) last week, the Chips and Science Act will soon become law. Besides the $54.2 billion provided to boost US-based chip manufacturing, the act reshapes US science policy in meaningful ways. NSF’s proposed budget... Read more…

Intel Is Opening up Its Chip Factories to Academia

October 6, 2022

Intel is opening up its fabs for academic institutions so researchers can get their hands on physical versions of its chips, with the end goal of boosting semic Read more…

AMD’s Genoa CPUs Offer Up to 96 5nm Cores Across 12 Chiplets

November 10, 2022

AMD’s fourth-generation Epyc processor line has arrived, starting with the “general-purpose” architecture, called “Genoa,” the successor to third-gen Eypc Milan, which debuted in March of last year. At a launch event held today in San Francisco, AMD announced the general availability of the latest Epyc CPUs with up to 96 TSMC 5nm Zen 4 cores... Read more…

AMD Previews 400 Gig Adaptive SmartNIC SOC at Hot Chips

August 24, 2022

Fresh from finalizing its acquisitions of FPGA provider Xilinx (Feb. 2022) and DPU provider Pensando (May 2022) ), AMD previewed what it calls a 400 Gig Adaptive smartNIC SOC yesterday at Hot Chips. It is another contender in the increasingly crowded and blurry smartNIC/DPU space where distinguishing between the two isn’t always easy. The motivation for these device types... Read more…

Google Program to Free Chips Boosts University Semiconductor Design

August 11, 2022

A Google-led program to design and manufacture chips for free is becoming popular among researchers and computer enthusiasts. The search giant's open silicon program is providing the tools for anyone to design chips, which then get manufactured. Google foots the entire bill, from a chip's conception to delivery of the final product in a user's hand. Google's... Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire