AMD Winds Down Year on a Sour Note

By Michael Feldman

December 14, 2007

It wasn’t easy being AMD for the last 12 months. By any measurement, 2007 was a miserable year for the company. Its stock, profits, market share and reputation have all taken a beating at a time when the company had expected to be using its quad-core processors to run rings around arch-rival Intel. Starting with losing its exclusive x86 relationship with Sun Microsystems in January to botching the Barcelona production ramp in November and December, nothing seemed to go right for AMD in 2007.

A lot of the company’s current troubles were set in motion by events in 2006, when Intel began its full court offensive against its smaller rival. By aggressively introducing new x86 product lines in the latter half of 2006 and by lowering chip prices, Intel started regaining market share, especially in the server space, which it had been losing to AMD since the introduction of the dual-core Opteron in 2005. In November 2006, Intel introduced the first x86 quad-core processors, beating AMD by almost a full year.

After AMD acquired ATI in October 2006, the effort of bringing the graphics chipmaker into the fold seemed to throw the chipmaker off its game. Now that it was delivering both CPUs and GPUs, AMD had to face the dual rivalry of Intel and NVIDIA. Over the ensuing 12 months, the company lost its x86 mojo.

When Sun Microsystems decided to form an alliance with Intel in January 2007 and introduce Xeon-based servers alongside its Opteron-based servers, AMD didn’t really suffer much in the deal, but it symbolized the loss of the Opteron’s momentum that had been in force since AMD first introduced the 64-bit x86 architecture in 2003. Soon after the Sun defection, Intel unveiled its 45nm processor technology, announcing its intent to introduce the new technology into commercial products later in 2007. The company made good on its plans, introducing the first 45nm Penryn chips in November, at a time when AMD was just starting to ship its first 65nm Opterons.

Meanwhile AMD’s plans to leverage ATI graphics chips for the high performance technical computing market were being outflanked by NVIDIA. Both companies announced initial GPGPU product offerings in 2006, but only NVIDIA brought a coherent product to market in 2007, with their CUDA programming environment and Tesla devices for high performance computing.

While Intel was flexing its quads in 2007, AMD kept itself busy nudging the ship date out for its initial “Barcelona” Opterons, which were originally scheduled to be delivered in mid-2007. When the quad Opterons were finally introduced in September, their relative speed and performance, compared to Intel’s latest chips, didn’t match the hype AMD had been feeding the industry over the past year. In the ensuing weeks, there were reports that AMD was having trouble meeting demand for the new quad-core processors. Rumors of chip yield problems abounded.

Which brings us to this week.

With the new Opteron chips shipping for three months, it was revealed that a bug related to the operation of the translation lookaside buffer (TLB) was found back in November by AMD. The bug (Erratum 298) affects the integrity of the data cache under certain circumstances. Apparently, this is most serious in commercial enterprise applications that use virtualization technology. At some point, the company decided to stop shipping Barcelona parts into the channel and to most system manufacturers. The end result is that the quad-core Opteron ramp-up won’t take place until Q1 2008, when they’ll have a fix for the bug in silicon. To add insult to injury, since the chips are no longer “generally available,” the Standard Performance Evaluation Corporation (SPEC) benchmarks associated with the quad-core Opteron will be removed from the SPEC website.

However, thousands of quad-core Opterons were delivered to high performance computing installations. Mercury Research reported 39 thousand quad-core Opteron chips were shipped in the third quarter. The Ranger supercomputer installation at the Texas Advanced Computing Center (TACC), alone, has 16 thousand of them. Since the TLB bug can be addressed with either a BIOS fix or a Linux kernel patch, presumably one or the other will be used at these sites to get around the problem. The BIOS fix comes with a 5 to 20 percent performance penalty, while the Linux workaround is said not to impact performance. In general though, most OEMs would be loathe to ship an uncertified patched OS kernel, but things work a little differently in big HPC deployments, where there’s a more intimate relationship between the system suppliers and the user.

According to the folks at TACC, the Opteron bug is not impacting the new Ranger supercomputer installation. The deployment there is now in full swing. Early users are being brought up on the system this week, with full production slated for next month. No word on which fix they used for the TLB problem.

Cray is moving forward on its Opteron-based supercomputer deployments as well. Apparently they chose to use the Linux workaround to avoid the TLB bug. “[O]ur early performance tests show that there is virtually no performance impact when using the Linux operating environment software,” stated Cray CEO Peter Ungaro. “In cooperation with AMD, we are running more tests and working closely with our early quad-core customers to determine delivery plans.”

Presumably, smaller HPC cluster deployments from vendors like HP, IBM, Dell, and Sun using the new quad Opterons with vanilla Linux won’t take place until new silicon is shipped in 2008. Bad news for HPC customers and AMD’s revenue. But hey, as anyone in the chipmaking business knows, errata happens. While I believe the new quad-core Opterons are great processors technically, overall product execution will determine the commercial success of the product. Because of the delayed production, it’s likely we won’t see volume deployments of quad-core Opteron-based clusters until next spring.

If all that weren’t bad enough, on Wednesday AMD filed a Material Impairments document with the SEC, stating that the company will take a “goodwill impairment charge” related to the ATI acquisition. In English, this means the company is admitting it paid too much for ATI and is going to take a substantial write-down in the current quarter to reflect what it believes is now the real value of the ATI business. The dollar value will be decided in the next few days.

On Thursday morning, AMD held a financial analysts meeting to explain the current state of the company and the future business strategy. In the meeting, the AMD execs portrayed the their situation with the “glass half full” metaphor, noting that from Q1 to Q3 the company has improved its average selling prices and gross margins, and moved ahead in the mobile market. They also managed a successful transition to the 65nm process technology and launched the 600 series GPUs. But how could “half full” apply to a year in which the company did not turn in a profitable quarter, had its stock lose more than a third of its value, and had its server market share drop from 26 percent to 13 percent? One might ask what a “glass quarter full” year would be like.

Much of the recent analyst and investor angst is, of course, about the debacle of the quad-core Barcelona launch. But AMD president and chief operating officer Dirk Meyer tried to play down the importance of the recent Opteron problems. He rejected the notion that the company needs to have the highest performing CPU in the market to ensure its success. “The prevailing wisdom is wrong,” declared Meyer. “The lion’s share of the market opportunity isn’t looking for the highest performing CPU. They’re looking for value, energy efficiency, great visual performance, and affordable Internet connectivity.”

An aggressive product map in 2008 will focus on what AMD considers the volume and revenue sweet spots: the desktop, notebook, mobile and HDTV markets. The company is going to devote particular attention to the commercial desktop and notebook space, where they see a large revenue opportunity. AMD plans to leverage its CPU and GPU chips to build platforms targeted at all these markets.

As far as the enterprise and HPC space goes, AMD claims it will not let Intel claim the high ground for the x86. “Hell no,” replied Mario Rivas, AMD executive vice president of the Computing Products Group, when asked if the company was going to cede the high end to its bigger rival. In mid-2008, AMD plans to introduce “Shanghai,” the 45nm generation of the quad-core Opteron. However, since the 65nm Opterons won’t ship in volume until late Q1 or early Q2, a new 45nm product line in Q3 would bump against the previous generation just as it was (hopefully) starting to generate some momentum. Maybe they have a pricing strategy that will allow both products to move forward in parallel for a while, but it seems like competition from Intel’s 45nm Xeon processors is going to force them to ditch the 65nm Opteron chips before too long — assuming they roll-out the 45nm technology on schedule.

Right now, AMD’s number one priority for 2008 is to return to profitability. Since they’ve identified the volume markets as the way to do that, their Opteron platform is bound to get relatively less attention. How that plays out for the high performance computing market should be one of the more interesting developments to watch in 2008.

—–

As always, comments about HPCwire are welcomed and encouraged. Write to me, Michael Feldman, at [email protected].

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