T-Platforms Places Bet on Switchless Supercomputing Interconnect

By Michael Feldman

October 27, 2010

Russian supercomputer maker T-Platforms is continuing its push into the elite end of the HPC market. On Monday, the company unveiled a joint venture with a group at the University of Heidelberg to develop a new ultra-fast interconnect for high-end supercomputing. The goal is to bring the technology to market in the form of an ASIC, which can be incorporated into a network interface controller (NIC) for HPC servers.

The technology, called EXTOLL, for Extreme Low Latency Interconnect, was born out of a research project led by Prof. Ulrich Brüning in the university’s Computing Architecture Group. The team there has created a working FPGA-based prototype for demonstration and software development purposes, but the final goal is to develop an ASIC for commercial production.

In a nutshell, EXTOLL is a switchless interconnect designed for ultra-low-latency, high bandwidth, and extreme scalability. The latter attribute is especially critical to the construction of large-scale HPC machines, which may contain tens of thousands of compute servers. According to Anton Korzh, a hardware architect at T-Platforms, the initial EXTOLL implementation will support up to 64 thousand nodes and at least a hundred cores per node. Even with the current generation of processors, that would allow systems to reach well into the multi-petaflop realm.

There is already an EXTOLL MPI software implementation in place, which has been developed and tested on the FPGA prototype. Since this technology is destined for petascale supercomputing, support for PGAS (Global Address Space) language environments is also in the works. Some support has to be baked into the operating system, and T-Platforms is planning to incorporate EXTOLL awareness into its own custom OS for HPC, ClustrX. In general, the proprietary software stack would be the biggest impediment to wider use of the technology.

Latency-wise, EXTOLL is aiming for sub-microsecond territory. The Heidelberg researchers believe they can achieve 500 nanoseconds (ns), NIC to NIC. No external switches are needed (switch logic is part of the NIC ASIC), and each hop in the network adds just 60 ns of delay. So even the worst cast latency for a 10,000 node supercomputer would be in the neighborhood of 3 microseconds — assuming a 3D torus network design, which is what T-Platforms has in mind for its implementation.

Bandwidth, too, is aimed at the upper end of the spectrum. Each of the six EXTOLL links on the NIC will be capable of transferring 120 gigabits/second, which works out to about 90 GB/sec for a single device. That’s about 2.5 times the speed of the current generation QDR InfiniBand, and puts it in the realm of the 160 GB/sec Cray Gemini interconnect used in the company’s newest XE6 “Baker” supercomputers.

The original EXTOLL design was based on the HyperTransport protocol, which would have effectively limited its use in supercomputing to AMD Opteron-based servers. T-Platforms persuaded the university researchers to incorporate PCI-Express (PCIe) support as well, so they could build hardware with Intel silicon. The current roadmap will include support for both PCIe 3.0 and HyperTransport 3.0. It’s worth noting that Cray’s future system interconnect, named “Aries,” will also support PCIe, and for exactly the same reason. The Aries technology is the follow-on to the Gemini interconnect, and will be used in the upcoming Cascade-class supercomputers.

While it’s a stretch to start comparing T-Platforms to Cray, the Russian vendor seems to be following the Cray model of layering a proprietary interconnect on top of commodity x86 parts for its top-of-the-line supercomputers. The big difference is that T-Platforms bought into the technology rather than developing it in-house. The arrangement between T-Platforms and the University of Heidelberg gives the company an equity position in the joint venture. In exchange, T-Platforms is investing an undisclosed sum in the project to help move the technology into commercial production.

The idea is for the supercomputer maker to get first dibs on the new interconnect so that it can be incorporated into a future blade product aimed at the upper end of the HPC market. T-Platforms intends to keep its InfiniBand-based blade, as well, for those systems that don’t require extreme scalability. According to Korzh, the company intends to move to a more modular blade design such that either interconnect technology (or perhaps even both) can be accommodated on the same basic motherboard.

The EXTOLL-based offering is already under development and is slated for launch in Q4 2011, when production of the NIC ASICs are scheduled to commence. The EXTOLL group is also interested in producing a stand-alone product that presumably would take the form of a PCIe-based network adapter that could be plugged into standard servers. That would make for a rather interesting setup for, say, a medium-sized supercomputer. Although the EXTOLL NICs are bound to be more expensive than their commodity InfiniBand or Ethernet brethren, the fact that one can do away with external switching could make for a compelling scale-out cluster model.

Getting the technology off the ground, however, is not going to come cheap. Typical costs for ASIC development alone can easily reach into the millions of dollars. Conveniently, T-Platforms recently announced an infusion of money from the state-run “Bank for Development and Foreign Economic Affairs” (Vneshekonombank), and although specific projects were not called out, the stated purpose of the investment was for “expanding T-Platforms’ supercomputing research and development efforts, along with providing support for the company’s expansion into the global HPC market.” These maneuverings appear designed as part of a concerted strategy to expand the Russian company’s reach into the European Union and beyond.

For the time being, though, Europe appears to be the primary target for T-Platforms’ supercomputing aspirations. The Russian (and former Soviet Union) high-end supercomputing market is too small, while the US one, represented mostly by DOE labs, DoD research centers, and NSF supercomputing centers, is under the implicit mandate to buy American. That said, a differentiated high-end offering from T-Platforms could shake up the positions of established European market players like Bull, Cray and IBM, in particular, and create a more diverse set of supercomputing choices than even the US enjoys.

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