OpenFabrics Alliance Weaves Its Story at SC09

By Nicole Hemsoth

November 15, 2009

We have developed something of a tradition at HPCwire in the weeks leading up to each year’s SC conference; we interview the chairman of the OpenFabrics Alliance (OFA). We realized five years ago the major impact that OFA’s free, open-source software stack could have on the HPC community, and we want to keep our readers updated on the work being done by OFA as well as the latest enhancements to the OpenFabrics Software stack.

Jim Ryan of Intel has been the OFA’s chair all these years, and our annual interview with Jim was as interesting as ever.

HPCwire: Good to talk with you again this year Jim. OFA members have been part of SCinet each year at SC. I suspect that is to allow exhibitors to show attendees the latest capabilities they’ve developed to leverage OpenFabrics Software on InfiniBand and Ethernet networks. What is being shown this year?

Jim Ryan: For the past four years, SCinet’s OpenFabrics Team has built networks at SC so exhibitors can demonstrate their products and research, including interconnects for servers, clusters, storage and file systems. This year we are setting records with the number of industry participants (18), the highest speed (IB 12X at 120Gbps) and the range of innovative applications that are on display.

OFA is collaborating with the HPC Advisory Council and together we are showing:

— Remote Desktop over InfiniBand (RDI) that enables live desktop sharing at high speeds between tens of participants.
— Direct Transport Compositor (DTC) that provides real-time rendering of a PSA Peugeot Citroen automotive CAD model in 2D/3D.
— New high-bandwidth MPI technology that takes advantage of 120Gbps data rates from a single server.

In addition to these demos, any connected exhibitors can demonstrate interoperability using OpenFabrics Software for Linux or Windows on the same IB network to exploit the full range of RDMA Application Services (R-DMAS). These include all the MPIs (e.g., Open, MVapich, Intel, HP), uDAPL, accelerated IP networking, Sockets using Oracle’s RDS, SRP and iSCSI for block storage, and file systems using any of NFS, Gluster, IBRIX, Lustre, and GPFS.

HPCwire: What else is on tap for OFA at SC09?

Ryan: We’ll be hosting a Birds of a Feather session at 5:30 p.m. on November 18 in Room PB251. At that time, we’ll have a very important announcement to make about a significant new member.

During the session, we’ll also announce the status, features and schedule for the next software releases: OFED 1.5 for Linux and WinOF 2.1 for Windows. Attendees will have the opportunity to ask OFA developers questions about many topics, such as improved routing for IB, routing options in Ethernet for clusters, the status of iWARP, and support in Linux environments for IEEE Datacenter Bridging on 10 Gigabit Ethernet for clustering. We’ll also be talking about our scalability road map, including plans for evolving OpenFabrics to be the interconnect software of choice for ExtremeScale.

On the exhibit floor, OFA is sharing Booth 137 with the InfiniBand Trade Association. There will be live demos and talks by our members as well as a joint presentation with the IBTA at 11:30 a.m. on November 17 at the Exhibitor Forum.

HPCwire: Can you tell our readers how OpenFabrics Software benefits end users?

Ryan: Well, the benefits must be pretty compelling because analysts estimate about 60 percent of all new HPC systems worldwide utilize OpenFabrics Software. We’re also seeing adoption in the enterprise where our software is being used to achieve high-speed server-to-storage communication and to implement unified fabrics on IB and low-latency Ethernet networks.

By linking applications to any of the RDMA services in the OpenFabrics Software, users can be assured they are getting the highest achievable data-transfer rates, the lowest achievable latencies (1 to 3 microseconds) and the most efficient computing (75 to 85 percent server utilization) for a full range of virtualization and enterprise datacenter applications, cloud computing, and simulation/modeling.

As an example, Oracle advertises RAC with IB as delivering a 10X improvement in application speed at half the hardware cost of traditional implementations. That’s made possible by OpenFabrics Software.

Users also gain the freedom of choice over a wide range of operating systems, computer architectures, and network and storage technologies. In each of these dimensions, we are seeing increased support from major vendors such as Dell, HP, IBM and Sun. All offer OpenFabrics Software with appropriate products. They also provide technical support.

HPCwire: Who are some of the members of the OpenFabrics Alliance?

Ryan: The primary network and interconnect vendors include Cisco, Mellanox, QLogic and Voltaire — all of which offer both Ethernet and InfiniBand products. The server vendors include Cray, HP, IBM and Sun. Storage is represented by DDN, LSI and NetApp, and databases by Oracle. Among the smaller members are Endace, System Fabric Works and Xsigo. Large users are represented by the DOE labs — Livermore, Los Alamos and Sandia. And for processors, we have Intel and AMD.

HPCwire: We hear a lot about Ethernet becoming the “converged network” that will be the only interconnect users need. How is OpenFabrics viewing this advocacy from certain vendors and what range of networks or fabrics are you supporting?

Ryan: As you know, InfiniBand has been the basis of our early work. Right from its inception 10 years ago, IB was conceived and architected as a unified fabric for all HPC and datacenter applications. In HPC, this vision is starting to take shape now. What people frequently overlook is that IB is architected in layers — application interfaces at the top, software in the middle and hardware at the bottom. Much of the value in IB is in the software which implements zero copy transfers, RDMA, APIs in the kernel, user space and reliable transports. So when we adopted iWARP a couple of years back that allowed the existing software to be integrated to provide an RDMA capability over Ethernet.

iWARP uses TCP/IP (that’s Layers 4 and 3) and Ethernet (Layer 2) as its underlying transport with the TCP/IP outboard on an adapter. This is excellent for users who need to maintain TCP/IP with store and forward routing of messages within their datacenters. At SC09, you can see iWARP demos at booths hosted by the Ethernet Alliance, Intel, Chelsio and others.

Now some organizations are starting to realize “Ethernet only” is all they need within their datacenters, particularly with the emergence of low-latency cut-through switches and DCB from the IEEE. Within OFA, some members have implemented the software from IB directly on both hardware-accelerated and legacy-Ethernet adapters and chipsets without any TCP or IP. It’s too early to tell right now how this will be accepted by vendors and users, but at SC you’ll be able to see a demo at the Ethernet Alliance booth.

HPCwire: What is OFA currently focused on?

Ryan: Let’s start with HPC. Reliable, efficient, scalable software for IB clusters that have 10 to 100,000 nodes is what our founding DOE Lab members are telling us they need in the next year or two. So that’s a high priority that will be of value to all users of OpenFabrics Software. It will be also become essential as IB speeds up to 40-80-120 gigabits with latencies below 1 microsecond.

I think we can all see in the next couple of years 10 Gigabit Ethernet prices for cluster connections (switch port and adapter/chip) dropping below $1,000. Then we will start to see an enormous percentage of clusters and datacenters migrate from their 1 Gigabit Ethernet networks to 10 gigabits. We believe that’s when RDMA over Ethernet, with or without TCP/IP, will hit its stride and converged or unified fabrics will become de rigeur. In the OFA, we are preparing both our Linux and Windows stacks to be the software of choice for this migration. OpenFabrics Software will be particularly valuable for organizations that need to support both legacy applications and RDMA applications.

Also in coming years, as we embark on extreme scalability in HPC, and virtualization and fabric unification penetrate deeper into datacenters, the OpenFabrics RDMA storage application services will become more important. It will enable traditional fibre channel SAN, iSCSI and NAS protocols to be used on the unified fabric, whether it be Ethernet or InfiniBand.

Jim, thanks for giving HPCwire’s readers such a detailed view into the OFA. You and your members have shown us the value of partnership and collaboration amongst vendors, developers and customers in an open-source community. See you in Portland at SC09 on November 15.

To keep up with the latest news from the OFA, goto www.openfabrics.org or join the OpenFabrics Alliance Facebook group.

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