Lenovo Unveils Warm Water Cooled ThinkSystem SD650 in Rampup to LRZ Install

By Tiffany Trader

February 22, 2018

This week Lenovo took the wraps off the ThinkSystem SD650 high-density server with third-generation direct water cooling technology developed in tandem with partner Leibniz Supercomputing Center (LRZ) in Germany. The servers are designed to operate using warm water, up to 45°C for general deployments and for special bid projects up to 50°C, lowering datacenter power consumption 30-40 percent compared to traditional cooling methods, according to Lenovo.

Nearly 6,500 of the ThinkSystems SD650s featuring Intel Xeon Platinum (Skylake) processors interconnected with Intel Omni-Path Architecture will be put into production at LRZ this year, providing the supercomputing center with 26.7 petaflops of peak performance, housed in a little over 100 racks.

The SuperMUC-NG supercomputer will be deployed with Lenovo’s new Lenovo Intelligent Computing Orchestrator (LiCO) and the Lenovo Energy Aware Runtime (EAR) software, a technology that dynamically optimizes system infrastructure power while applications are running.

Lenovo’s Scott Tease holding a ThinkSystem SD650 server

“Pretty much all the investments that we made to get to exascale LRZ is taking advantage of in this bid we won with them,” said Scott Tease, executive director, HPC and AI at Lenovo in an on-site briefing at Lenovo’s headquarters in Morrisville, North Carolina, last week. “We will start building systems and start shipping them in March; the floor will be ready by the end of April, and move-in starts in early May. We’ll be ready to do acceptance in September with final customer acceptance in November.”

The direct-water cooled design of the SD650 enables 85-90 percent heat recovery; the rest can easily be managed by a standard computer room air conditioner. The hot water coming off the servers can be recycled to warm buildings in the winter, as LRZ does with its petascale SuperMUC cluster, but the technology developed by Lenovo for SuperMUC-NG actually transforms that heat energy back into cooling for networking and storage components.

The endothermic magic trick only works with “high quality heat,” Lenovo thermal engineer Vinod Kamath told us, so LRZ’s SD650 servers were designed to be able to consume 50°C inlet temperatures. Water is piped out of the servers at 58-60°C depending on workload and sent through an adsorption chiller, where it is converted to chilled 20°C water suitable for cooling storage and networking components.

If you’re using chilled water to cool servers you can’t really take advantage of the economics of the adsorption chiller. With 60°C inlet water, the efficiency of Lenovo’s adsorption chiller is about 60 percent. If your energy source has a higher temperature, say 80-90°C then the extraction is even more efficient, but 60°C is good enough to realize significant savings.

Adsorption chilling will be applied to half the nodes of the next-gen LRZ install, generating about 600 kilowatts of chilled water capacity. This translates into more than 100,000 Euros a year in saved energy cost at the European site, where the rate for energy is about 16-18 Eurocents per kilowatt-hour (roughly 2-3 times the cost for similar sites in the United States). Lenovo claims a 45-50 percent energy savings with the endothermic reaction versus a traditional compressor, dropping the datacenter PUE from 1.6 to less than 1.1.

click to enlarge

The cooling solution can be traced back to 2012, when IBM (Lenovo acquired IBM’s x86 server business in 2014) was approached by LRZ to develop a system that was both powerful and extremely energy efficient. The first production implementation to come out of the partnership was the 9,200 node SuperMUC at LRZ, that achieved a number four ranking on the June 2012 Top500 list. The custom motherboard, developed with Intel, was cooled by water piped over compute and memory and back out of the system. LRZ used the hot water coming out of the system to heat parts of their building, which offset some of their overall energy costs.

The partnership also led to the deployment of the CooLMUC-2 cluster at LRZ in 2016. That system was the prototype for the next-gen LRZ cooling solution; it uses hot outlet water to drive adsorption chillers that generate refrigerated water, which is then used to cool some of the cluster’s disk storage systems.

“When we started doing this it was all about power cost,” said Tease. “It was all about datacenter optimization. Those things are still important, but we’re starting to see people recognize that water will allow them to do things that air can’t. I can do special processors that I can’t do with air; I can achieve densities that in the future I can’t do with air. We are really excited that we’ve got such a unique design, what we believe is an industry-leading design point as the market is coming to where we’ve been.”

The Lenovo ThinkSystem SD650 dual-node tray is designed for high-performance computing (HPC), large-scale cloud, and heavy simulations. One 6U NeXtScale n1200 enclosure houses up to six of these trays, accommodating a total of 12 SD650 compute nodes, 24 Intel Xeon Scalable Processors, 9.2TB of memory, 24 SFF SSDs or 12 SFF NVMe drives, and 24 M.2 boot drives.

Front view of the n1200 DWC enclosure. Source: Lenovo

The SD650 HPC servers have no system fans (except on the power supplies at the back of the rack), and operate at lower temperatures when compared to standard air-cooled systems. Chillers are not needed for most customers, which translates into further savings and a lower total cost of ownership. The new server supports high-speed EDR InfiniBand and Omni-Path fabrics as well as standard SSDs, NVMe SSDs, and M.2 boot SSDs.

In demoing the SD650, Kamath showed how the water supply comes in through the 6U NeXtScale n1200 chassis and flows into the servers. “We have a calibrated flow split between the processor and the memory to tune the heat transfer,” he said. “We recognize that networking devices are power hungry now and will be more so in the future, so the water that splits to the memory is coupled to a drive, an NVMe or SSD, and coupled to a network device, like ConnectX-5 or OPA, and then the water flows and connects back to conduction point.”

Two Lenovo ThinkSystem SD650 servers on the compute tray that provides water cooling. Source: Lenovo

Lenovo designed the system with special attention to the next-generation memory technologies. Each server has 12 DIMM slots for truDDR4 memory but there are actually 16 slots total. Four have been reserved for 3D-XPoint (also known as Apache Pass or AEP memory). The cooling system is able to extract 10 watts on standard DIMMs, and for 3D XPoint and other higher-powered memory future designs, they’ll have two water lines going through a DIMM that can consume 18 watts. Lenovo also provides a handy DIMM removal tool making it easy to swap out memory.

Lenovo has been picking up major HPC system awards in Europe since acquiring IBM’s x86 business three and a half years ago. It has the fastest supercomputer in Spain, Italy, Denmark, Norway, Australia, Canada, and soon in Germany with LRZ. It has also been making in-roads with its warm water cooling solutions. In addition to its systems at LRZ, it has warm water HPC installations at Peking University (first ever in China), India Space Administration (first ever in India), and a multi-university system in Norway.

Liquid cooling is becoming mainstream in HPC, especially in environments where constraints on space boost density requirements or in expensive energy zones. Lenovo tells customers that when it comes to electricity prices, anything over 15 cents per kilowatt hour will provide a return on investment within one year. Another benefit of removing more heat is that CPUs can run in “turbo” mode nonstop, which can squeeze an additional 10 percent performance from them.

The SD650 is managed by Lenovo Intelligent Computing Orchestrator (LiCO), a management suite with an intuitive GUI that supports management of large HPC cluster resources and accelerates development of AI applications. LiCO works with the most common AI frameworks, including TensorFlow, Caffe and Microsoft CNTK.

SD650 system architectural block diagram. Source: Lenovo
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