SGI Revamps Altix ICE Design

By Michael Feldman

July 5, 2011

For the past five years, SGI’s Altix ICE platform has been the company’s bread and butter HPC cluster offering. That trend looks to continue as they gear up for their fifth generation design. But this iteration of Altix ICE, codenamed “Carlsbad 3,” is more than just a processor and InfiniBand refresh.

According to SGI marketing VP Bill Mannel, the first four generations of Altix ICE were variations on a theme, using essentially the same blade infrastructure. As Intel or AMD came up with new processors and Mellanox rolled out faster InfiniBand parts, SGI slapped them into the Altix ICE lineup. But Carlsbad 3 has been redesigned from scratch, says Mannel.

Like many HPC blade vendors, SGI is going after the three big attributes customers are clamoring for: higher density, lower cost, and more efficient cooling. Just riding the Moore’s Law curve for microprocessors gets you halfway there for the first two attributes, but better cooling required some re-engineering on SGI’s part.

The previous versions of Altix ICE relied on either standard air cooling or liquid cooling in the form of water cooled doors bolted onto the enclosures. But like a lot of denser blades, SGI has added the option for a cold plate, where the liquid runs against the hottest components of the blade. That design also allow for warm water cooling, where the water temperature can run up to a tepid 30 degrees Celsius (86F), and which can be cooled via a liquid-to-air exchange rather than a power-sucking chiller. Although this is a new feature for the Altix ICE, the cooling technology is derived from the Rackable ICE Cube container that SGI offers today.

Mannel says the 30C limit for warm water cooling is showing up in more RFPs, and is becoming more accepted as datacenters, HPC or otherwise, are forced to cram more compute capacity into the same building. For supercomputing setups, this can be especially useful for customers who want to build their systems with extra-hot processors like high-bin (high GHz) x86 parts and big wattage accelerators like GPUs.

Speaking of which, the new ICE machines will sport the new Sandy Bridge EP Xeons, Intel’s upcoming CPU offering for dual-socket servers. As far as accelerators go, at the very least SGI will be offering NVIDIA’s latest GPUs, either the standard M2090 Tesla modules or the related X2090 designed for extra-dense blade setups. According to Mannel, the final design is being worked out this month.

And although Intel’s MIC (Many Integrated Core) Knights Corner coprocessor won’t be arriving until later in 2012, it’s a pretty sure bet that the Altix ICE will adopt them when the Intel releases the commercial product. At the International Supercomputing Conference last month, SGI was one Intel’s system partners cheerleading the MIC development effort. (Currently SGI offers a MIC development system based on the Rackable H4002 server and the Knights Ferry part.) In fact, MIC is apt to show up across multiple SGI HPC offerings next year. “We do intend to include it into our system plans going forward,” says Mannel.

InfiniBand, too, got an upgrade in the Carlsbad 3 design. The default interconnect will be Mellanox FDR InfiniBand, but since SGI is using mezzanine cards for the network I/O, the customer could opt for 10GbE too. The rational behind mezzanine cards is that they are cheaper than using discrete network adapters and more flexible than hardwiring each blade with as specific interconnect by plopping the network silicon directly onto the motherboard. Also, when EDR InfiniBand goes into production, customers should be able to just swap the FDR cards with the EDR version, leaving the compute blade as is.

Also for increased flexibility, the SGI engineers went to a power shelf design. In the current ICE setup, the power supplies are inside the blade enclosure; with the new design, they will be in separate units. That allows customers to add (or subtract) power supplies as needed, as for example, when installing more blades or hooking up accelerators.

Density wise, Carlsbad 3 will offer two configurations, each more compact than the current Altix ICE 8400. In the standard configuration, the new ICE will up the blade count from 16 to 18 per 10U enclosure. That’s 12 percent more compute per rack unit. And since the Sandy Bridge chips will provide up to 2 more cores per socket than the 6-core Westmere EP silicon in the 8400, the compute density gets another 33 percent boost.

Also, instead of coming in 24-inch wide racks, the engineers have squished them into standard 19-inch boxes. That will enable the new machines to more easily co-exist with vanilla datacenter gear.

The other density innovation is their M-Rack configuration. Basically it’s a double density setup that is able to fit twice as many blades (36) into an enclosure by sandwiching two motherboards into a blade slot — what they call the Gemini Twin blade. The M-Rack essentially flips two standard racks 90 degrees and pushes them together, squeezing out the space that would have been the hot aisle (and compensating with the type of water cooling discussed above.) The M-rack is nearly twice as dense as the new standard density offering and nearly 3 times as dense as the current Altix ICE 8400.

The new ICE machines are scheduled to start shipping in December and according to Mannel they already have some customers in the pipeline. Although he didn’t say, one of them could be NASA, which recently upgraded their Altix ICE Pleiades supercomputer with additional hardware. That system is now in the petaflop club and sits at number 7 on the TOP500 list.

SGI’s biggest supercomputer to date, Pleiades represents a microcosm Altix ICE history. The system was originally installed in 2008 and has been upgraded on a continuous basis ever since. Through the aggregation of Altix ICE hardware of various generations, it now contains a mixture Harpertown, Nehalem and Westmere Xeon processors and a combination of DDR and QDR InfiniBand networks.

Pleiades may end up as a 10 petaflop system as early as next year, as was the original intention of NASA. If so, that is almost certain to be accomplished with the upcoming Carlsbad 3 blades, most likely souped up with accelerators. And given SGI’s and NASA’s penchant for Intel silicon, those accelerators could very well be MIC coprocessors.

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!

ARM, Fujitsu Targeting Open-source Software for Power Efficiency in 2-nm Chip

July 19, 2024

Fujitsu and ARM are relying on open-source software to bring power efficiency to an air-cooled supercomputing chip that will ship in 2027. Monaka chip, which will be made using the 2-nanometer process, is based on the Read more…

SCALEing the CUDA Castle

July 18, 2024

In a previous article, HPCwire has reported on a way in which AMD can get across the CUDA moat that protects the Nvidia CUDA castle (at least for PyTorch AI projects.). Other tools have joined the CUDA castle siege. AMD Read more…

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…

SCALEing the CUDA Castle

July 18, 2024

In a previous article, HPCwire has reported on a way in which AMD can get across the CUDA moat that protects the Nvidia CUDA castle (at least for PyTorch AI pro 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…

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…

Shutterstock_1687123447

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

Contributors

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…

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…

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…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire