AMD Unveils New Strategy for Server Silicon

By Michael Feldman

February 7, 2012

AMD is plotting a relatively conservative roadmap for its Opteron CPUs over the next year or two, even as it preps its heterogenous computing technology for the big leap into the server arena. At the company’s 2012 Financial Analyst Day last week, AMD execs re-pledged their commitment to the server market and outlined a strategy that puts less emphasis on high performance cores and design complexity and more on power efficiency and building SoC products tailored to specific datacenter workloads.

In the near term though, it will be very much business as usual for the Opteron line. The big news (or non-news) is that AMD will not follow the top-of-the-line 16-core Interlagos chip with a 20-core successor — the so-called “Terramar” CPU. Instead, the company will offer “Abu Dhabi,” which, like its predecessor, tops out at 16 cores. It also uses the same processor technology (32nm) and offers the same memory support (quad-channel DDR3) There is no support for PCIe Gen 3, which was skipped for this go-around with the rationale that the newer, faster bus interface won’t be needed “until the market is better positioned for wide adoption of that very high-end technology.”

Abu Dhabi and the other next-generation Opterons (“Seoul” and “Delhi”) will be based on a new core architecture, known as “Piledriver,” and are scheduled to be launched in the second half of 2012. Rather than a complete redesign, Piledriver appears to be a tweak of the modular CPU design AMD pioneered with Bulldozer last year. The architectural enhancements include new ISA extensions and improved IPC. Essentially AMD is doing two tocks (microarchitecture redesign) in a row, with no intervening tick (process technology shrink).

The chipmaker’s conservative Opteron strategy may reflect some new thinking there. According to AMD’s new CTO, Mark Papermaster, in the past the company has put too much effort squeezing the last ounce of performance out of the cores, using extra design complexity to compensate for second place in semiconductor manufacturing. “We’ve gone after that last two to three percent performance, and historically it’s led to a longer development cycle,” said Papermaster. According to him, the new focus is on time-to-market and using hardware-software codesign to deliver application platforms, rather than just silicon.

Lisa Su, Senior Vice President and General Manager or AMD’s Global Business Units, said the decision not to scale up the core count on the next-generation Opterons was the result of customer feedback. According to her, rather than wanting more cores, their server clients were just interested upgraded Opteron parts that could be plugged into the existing G34 and C32 sockets. In any case, since AMD has no fab partner that is ready to move to a sub-32nm process, there really wouldn’t be additional die space available for more cores, caches, and bigger memory controllers without a much more drastic microarchitecture design.

According to Su, the new Piledriver cores will deliver more performance at the same TDP, although, at this stage, AMD is not offering any numbers that would shed light on those improvements. If the company can eke out some additional FLOPS from the Piledriver cores, along with some interesting ISA extensions, that’s probably their best shot at competing against Intel’s Sandy Bridge Xeon CPUs (E5 series), which are built on 22nm technology. The Xeon E5 CPUs are already installed in a number of top supercomputers, although the chips are not officially launched yet.
 
Delphi, in case you were wondering, is the successor to the not-yet-released Zurich CPU, which is based on the current generation Bulldozer core. Zurich is slated for release in the first half of 2012. These 1P processors will inhabit a new socket known as AM3+, code-named “Jakarta.” They are being targeted to light-weight web serving and related microserver space, which Intel has recently made a play for with its low-power Xeons and higher-end Atom chips.

Beyond Piledriver, is the Steamroller architecture, another modular CPU design that promises greater parallelism, which could mean either more cores or simultaneous multi-threading or both. After Steamroller come “Excavator,” a microarchitecture that focuses on greater performance. No dates were attached to either of these designs but 2013 and 2014, respectively, would be likely timeframes.

Of course, the other side of AMD is their GPU portfolio. But it’s notable that none of the product talk during the Financial Analyst Day mentioned the company’s FireStream offerings, the company’s discrete GPU accelerators aimed at high performance computing. In the face of nearly complete dominance of NVIDIA’s Tesla products for this space, it’s likely that AMD has ceded this market to its rival, at least for the time being.

Where AMD has a clear advantage is its ability to marry its CPU and GPU logic onto integrated heterogeneous chips, which they call APUs (accelerated processing units). All the APUs the company has developed to date have been targeted to client devices — desktops, notebooks and soon tablets. Not surprisingly, company execs devoted much attention to the APU client roadmap during the Analyst Day. But there was also a fair amount of discussion about migrating APU designs into the server space.

In particular, AMD sees custom datacenter workloads in areas like multimedia web serving, search engine processing, visual rendering, high performance computing as an opportunity for GPU acceleration in their heterogeneous computing platforms. One aspect to this is that they intend to be able to build SoC products in a modular way that combines x86 cores with their GPU designs, and do it in such a way as to tailor different chip designs to different workloads. AMD is even willing to incorporate third-party IP blocks into these SoCs, for example, fixed-function cores aimed at very specific types of processing like codec encoding/decoding.

The logic behind this strategy is that because many of these workloads are feeding the boom in web-connected mobile devices, there is a huge and rapidly growing market for such server infrastructure. “You’re not just talking about racks and racks of servers that just care about power and performance,” said Su, “you’re talking about specialty workloads, and it actually will fragment the server market a bit.”

Citing IDC numbers, AMD points to projected compounded annual growth rates of 15 percent, for cloud-based web applications; 13 percent, for virtualized workloads; and 7.3 percent, for high performance computing, over the next three years. Whether those numbers pan out as forecast and are enough to support volume production of specialized SoCs remains to be seen, but AMD doesn’t want to left with its one-trick Opteron pony if the server market starts to fragment.

Executing on that strategy is going to be the principle challenge for AMD. In the short-term, it has to find a way to get its server market share above the single-digit level — 5 to 7 percent in 2011, by most estimates — on the merits of its Opteron line. But the more difficult task ahead will be moving its heterogeneous technology into the datacenter. Although AMD has more of the pieces in place than its competitors, the heterogeneous waters here are uncharted. Nimbleness will be well-rewarded.

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