Marvell Talks up ThunderX3 and Arm Server Roadmap

By John Russell

March 17, 2020

Marvell yesterday released more details about ThunderX3, its next-gen 7nm Arm-based microprocessor, codenamed “Triton,” which it says is now sampling and will be broadly available later in 2020. The new chip will feature up to 96 Arm v8.3+ cores and support 4 threads per core thus delivering up 384 threads per socket. While many details of ThunderX3’s architecture were not disclosed, Marvell says more information will be forthcoming over the next few months. Marvell also took the opportunity to issue a barrage of performance advantage claims over both Intel and AMD CPUs.

The recent rise of Arm CPUs in servers targeting HPC and the cloud is noteworthy. Arm has long been a force in embedded applications, leveraging its low power consumption attributes. Now, advancing chip features, available silicon, an emerging accelerator strategy, and a significantly expanded ecosystem are invigorating Arm’s server aspirations.

Gopal Hegde, Marvell

Marvell, through its acquisition of Cavium (announced 2016, completed 2018) is finally starting to enjoy growing success in servers with the ThunderX CPU line. ThunderX2, for example, is being used in high-profile supercomputing projects at Sandia Laboratory (Astra), Los Alamos National Laboratory, GW4 (the Met Office), and France’s CEA. In the cloud, Microsoft Azure now uses ThunderX2-based clusters for internal purposes and Marvell says it has deals with 20 other hyperscalers.

Of course Marvell isn’t alone. Fujitsu’s A64FX Arm CPU will power Japan’s “Fugaku” supercomputer to be deployed at RIKEN in 2021, and in November Cray (now HPE) announced a collaboration with Fujitsu to bring out A64FX-based systems.

All things considered, the Arm camp –  once thought of as a long shot in mainstream servers let alone high-end HPC – is making steady headway into server markets. In a pre-briefing on the forthcoming ThunderX3 with HPCwire, Gopal Hegde vice president and GM, server processors, Marvell, argued that now that the heavy lifting is done, Marvell’s Arm’s server chip design is inherently better than x86 architecture because it doesn’t need to support legacy architecture or so many diverse device types.

“Intel designed its cores for use in [systems] from laptops and desktops all the way to servers. It’s not optimized for servers. We have no x86 legacy, like 32-bit support and things like that,” said Hegde. “We are able to optimize our code, and our core area is significantly smaller [as a result]. Just to give you an idea, in the previous generation, if you look at ThunderX2, compared to AMD or Skylake, for the same process node technology [we get] roughly 20% to 25% smaller die area. That translates into lower power. When we move to 7nm with ThunderX3, our core compared to AMD Rome’s 7nm is roughly 30% smaller.”

The slides below summarize ThunderX3’s specs, Marvell’s general pitch for ThunderX in advanced computing, and its processor portfolio:

 

It may be useful to briefly describe Marvell. Founded in 1995, FY20 revenue was $2.7 billion. The company has roughly 5000 employees. Its roots are in storage technology, but the product portfolio and markets served have expanded over the years. Marvell now has three main businesses – processors, networking, and storage, and it bills itself as the largest supplier of Arm server chips.

“Cavium was in the processor business for almost 15 years. Marvell has been shipping Armada (low power SoC) for a similar amount of time. So together, we have shipped over hundreds of millions of CPUs over the years. These are multicore CPUs ranging from two cores all the way up to 48 cores are now, even higher (96-core ThunderX3) soon,” said Hegde

“Octeon and Octeon Fusion are products used in wireless 5G infrastructure. We announced design wins with Samsung and Nokia about a week ago. Octeon processors are [also] very widely used in the embedded market and constitute a pretty significant part of the 2.7 billion in revenue that we did last year. Today we exclusively develop Arm-based processor products,” he said.

The server-oriented ThunderX line targets HPC, the cloud, and as Marvell puts it, “Arm native applications at cloud and edge.” Hegde contends that Intel’s struggles with process and resulting low core counts have created an opportunity for making gains in single and multi-threaded performance, while AMD’s multi-die-on-a-chip, or chiplet, approach necessarily introduces latency. ThunderX3, said Hegde, is designed to exploit those.

Given small die area and the lack of legacy x86 overhead, contends Hegde, it is possible to leverage Arm architecture for lower power consumption, lower cost, and added functionality “into the same monolithic die.” This gives ThunderX3, he argues, improved instruction per cycle (IPC) performance, better thermal design power (DTP), and “really good memory latency and memory bandwidth. “If you look at the ThunderXs, from that standpoint, you get the best of both worlds. You don’t have to sacrifice core count like, x86 Intel, and you don’t have to sacrifice memory latency like AMD.”

It’s true ThunderX3 still doesn’t have Arm Scalable Vector Extension (SVE) which Fujitsu’s A64FX has. Hegde said “It is going to be available in a later processor. The challenge is that compilers needed to take advantage of SVP are still under development. So we’re actually pretty happy to be providing systems so that the compiler toolchain can evolve.”

Below are a few workload and performance comparison slides Hegde presented

Until the fairly recently, the state of the Arm ecosystem – tools, supported OSs, adapter cards, etc. – has been a source of concern among would-be Arm users. That does seem to be changing. One prominent example of change is Nvidia’s decision last summer to support Arm as its accelerator strategy emerged and firmed up.

Hegde noted, “When I started in 2014 [with Cavium], we had two partners. Over the last six years, we have built a very broad ecosystem of over 100 partners across commercial, open source, and industry standard partners. A lot of these are driven through contracts and contracts to platforms that are in the ecosystem. Today there are thousands of ThunderX and ThunderX2 platforms in the ecosystem today.

“Not only do we work with several OEMs and ODMs to deliver platforms, but also have full tools and full collaterals. Pretty much all the operating systems are supported on ThunderX today, ranging from, Red Hat, SuSe, Oracle Linux, to Microsoft Windows, VMware, and some of the free community-based operating systems like Centos, FreeBSD, all the way up to middleware. HPC has been a special area of focus for us. And the number of partners in that space has more than doubled over last 12 months, and of course, cloud and now also in the edge computing,” he said.

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