Author Archive
People are talking about the 40-watt ACP Six-Core AMD Opteron™ EE Processor
Posted by Phil Hughes in 4:04 PM
(6 votes, average: 4.83 out of 5)
Loading ...Yesterday we launched our latest energy-sipping server processor, the 40-watt ACP, Six-Core AMD Opteron™ EE processor. For those of you scoring at home, 40W ACP / 6 cores = ~6.67W per core.
In case you didn’t have time to read the enormous volume of coverage from Yesterday’s news, we at AMD’s Global PR headquarters here in Austin, TX are happy to provide you with the following summary:
Blogger’s note: Some of you may remember back in 2004 we brought to market a single-core, max TDP 30-watt processor. Very much ahead of the curve as the article suggests.
Here’s some additional info on DDR2 vs. DDR3 memory pricing for servers. Latest data shows customers still need to pay a premium for DDR3.
Here's a chart that shows how AMD does not compromise features in order to save power
“Advanced Micro Devices is also hoping to appeal to cost-conscious data center managers with a new line of low-power, six-core “Istanbul” Opteron CPUs. The 40-watt models join the 75- and 55-watt versions already on the market.” Good Morning Silicon Valley
The above doesn’t even include the 100+ tweets and counting I’ve already seen on this since yesterday. Only a few of the tweets are from me. So yes, people other than me are tweeting about this amazing engineering and manufacturing feat that is the 40-watt ACP Six-Core AMD Opteron EE processor.
Phil Hughes is a senior public relations manager at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.
Is it time to stop trying to define the Cloud?
Posted by Phil Hughes in 1:33 PM
“Either you stand for something or you don’t”
Jack Ryan from the Tom Clancy novel Clear and Present Danger
Cloud Computing is real and seems to be growing rapidly. Cloud providers need the most energy-efficient servers they can get without compromising the features and reliability of traditional rack servers. That aspect of Cloud Computing seems to be, well, the least cloudy. It’s why AMD added the 40-watt ACP Quad-Core AMD Opteron™ EE processors to our server lineup earlier this year and plans to soon offer 40-watt ACP Six-Core AMD Opteron™ EE processors.
Since we launched our EE processors, I’ve tried to gain a better understanding of what Cloud Computing really is. Now, I’m not sure why, but that over-simplistic Jack Ryan quote kept running through my head as I recently read the latest f5 networks survey results on Cloud Computing. It asked 250 IT professionals what they thought of SIX possible Cloud Computing definitions. Before reading the results, I said to myself: Either you know what the Cloud is or you don’t. It should just be that simple.
Not so fast. There appears to be a thought leadership race among analysts, service providers and numerous other industry pundits to be the one who arrives first at the accepted Cloud Computing definition.
This “define the cloud” fascination seems to be reaching new levels. This week alone, I found more than 100 articles that discussed the topic of defining Cloud Computing on Google News. Does this mean we are near the “tipping point” on coming to a consensus, or are we further away than ever?
India’s CXO Today just posted an Cloud Computing interview with an SAP senior executive, Simon Dale. Dale said “For now it is still a marketing definition with nobody having come out with a proper and complete description of cloud computing.” Great observation and it would be nice if the industry could actually land on even a single “marketing” definition of Cloud Computing.
The good news is that even without a widely-accepted definition, interest and investments in Cloud computing are undeterred, looking at f5’s survey results. More than 80 percent of the 250 IT professionals surveyed are now in at least trial stages of Cloud deployments, and more than half have deployed a public Cloud environment.
Now as a PR guy, I’m all for the discussion about defining Cloud Computing to rage on. Particularly if it drives attention to the fact that AMD is working hard to shave every watt it can off its processors without compromising the features such as AMD-V™ or AMD-P. While the definition remains cloudy, we continue to work to keep the benefits to customers clear.
Phil Hughes is a senior public relations manager at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.
Boom or Bust
Posted by Phil Hughes in 4:13 PM
This is the final chapter of my blog looking back on the first year of the AMD OpteronTM processor.
If you’re a fan of professional sports, you know that the stretch from April until the end of June is draft time for the big four North American leagues -- NFL, MLB, NBA and NHL. Inevitably during these months, the sports media revisit past drafts and look at some of the biggest draft busts and successes. Take an example from the NFL -- Ryan Leaf (2nd pick, 1st round, 1998) and Tom Brady (199th pick, 6th round 2000) -- epitomize the concepts of bust (Leaf) and boom (Brady.)
Contrast this to the world of microprocessors. The AMD Opteron processor entered the market in April 2003 as a solid, but unproven, option during a time when the industry had been trying to determine if the 64-bit Itanium processor was more hype than substance.
I noted in my first blog that the AMD Opteron processor was met with some skepticism early on due to a lack of initial OEM support. But by the first half of 2004, the AMD Opteron processor was well on its way to being the dark-horse in x86 servers. Sun began to launch the servers it announced late in 2003 and the market was seeing some impressive performance on those systems. In February, the competition did us a big favor when they validated our approach to 64-bit computing by emulating the AMD64 instruction set extensions. In fact, it kicked off the first of several instances in that period where the competition took a page from the AMD playbook (multi-core, processor model numbers and Direct Connect Architecture are just a few more examples of them following the AMD lead.) The competition announcing its plans to incorporate AMD64 extensions served notice to the industry that AMD was ahead of the curve in understanding the needs of enterprise customers.
We had fun at IDF in Feb. 2004 reminding the world who made 64-bit migration easy.
A big helping of “Swordfish”
In the first few months of 2004, the momentum behind the AMD Opteron processor was fast and furious. The teams that supported the AMD Opteron processor were extremely busy but no one was complaining. We all had a lot of sweat equity invested in the AMD Opteron processor and we were now seeing the payoff. While IDF was going on, several of us were working on a project code-name “Swordfish”. “Swordfish” was the codename for our upcoming server launch with HP and it represented probably the biggest milestone for the AMD Opteron processor. Here’s why;
- Most industry watchers assumed HP would probably be the least likely OEM to come on board because of their investment in Itanium;
- HP didn’t make a small commitment to AMD, much like Sun, they went big right out of the gate;
- As the industry-leader, HP’s broad portfolio of AMD servers helped drive the discussion away from AMD being seen as a niche player in HPC;
- HP had a huge footprint among Wall Street IT, which was a critical customer segment for AMD in order to increase our market share.
Slide from HP's launch presentation on February 24, 2004
Rookie of the Year
In April 2004, we went back to NYC to celebrate the one-year anniversary of the launch of the AMD Opteron processor. While our event didn’t quite have the production value that our launch did a year earlier, we opted to go for substance over style. What a stark contrast from a year earlier. IBM was of course back again but joined this time by our new friends from Sun and HP along with an end-user from Verisign.
We had our cake, and ate it too at the AMD Opteron one-year anniversary celebration in NYC
One final thought: at its launch in April 2003, the AMD Opteron processor delivered some unique innovations to an x86 server market hungry for choice. The AMD Opteron processor initially fell into that category of a draft pick that had loads of potential, but not quite a first rounder yet. However, looking back at some of the milestones AMD achieved with the AMD Opteron processor from April 2003 to April 2004, it probably was a clear-cut “Rookie of the Year,” in the x86 server business. Particularly if you read some of the following….
“AMD is viewed as a credible tier-one server processor supplier, which certainly wasn’t the case a year ago,” said Dean McCarron, principal analyst at Mercury Research. InfoWorld, April 19, 2004
The Opteron processor has vaulted AMD to the front lines of the competition between IBM, HP and Sun Microsystems--three of the top four server makers. CNET, April 19, 2004
Overall, AMD shattered industry expectations for Opteron’s first year. The fact that the company’s product is even discussed as a serious Intel alternative in the server market is a major accomplishment. The Register, April 23, 2004
There you have it, some great quotes that sum up the first year of AMD Opteron. Perhaps in another year or two, you will see another blog from me about the first year of our Six-Core AMD Opteron processor, also known as “Istanbul”. Thanks for reading.
Phil Hughes is a senior public relations manager at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.
Recap of early “Istanbul” Reviews
Posted by Phil Hughes in 8:51 AM
(4 votes, average: 4.00 out of 5)On June 1 we launched our Six-Core AMD Opteron™ processor, formerly code-named “Istanbul”. Here are some of the early highlights from some of the industry’s most-respected hardware reviewers.
AnandTech: AMD’s Six-Core Opteron 2435
“It is out of the scope of this article, but it is clear that even if the CPUs cost the same, the AMD based server will be less costly.”
“Our power consumption data is preliminary. We really have to double-check all the power data. Very roughly, we find that the Opteron 2435 machine consumes about 35-45W less than the Xeon X5570. On a total of slightly more than 300W, that is about 10 to 15%. Idle power seems to be slightly in favor of the Xeon “Nehalem”.
“There are two types of applications where we feel that the AMD six-core deserves your attention: decision support databases and virtualization. . . . Of course, performance-critical applications are by definition more demanding when it comes to processing power. That is exactly what vApus Mark I measures: how well do performance-critical applications perform when they are virtualized? This is a relatively “new” market where the AMD 2435 shines. The new Opteron 2435 at 2.6 GHz was a pleasant surprise on vApus Mark I: it keeps up with more expensive Xeons on ESX 3.5 update 4 while consuming less, and offers a competitive performance/watt and performance/price ratio on vSphere 4. The six-core Opteron is about 11 to 30% slower on vSphere 4 than the 2.93 GHz Xeon X5570 but the overall cost of the Istanbul platform is significantly lower (DDR-2 versus DDR-3) and the 2.6 GHz 2435 consumes less power in a virtualized environment (*). On the condition that you optimize your hypervisor well to take advantage of the six cores (cell size is for example one critical optimization), we feel that the six-core Opteron is a worthy opponent for the Xeon “Nehalem” in this market. We tested only the 2435 versus the X55xx series.”
(*) Virtualized servers do not run idle very often
Tech Report: “AMD’s ‘Istanbul’ six-core Opteron processors / We crack open a six-pack”
AMD Executing Early: “AMD had originally planned to introduce this product in the October time frame, but the first spin of Istanbul silicon came back solid, so the firm pulled the launch forward into June.
“The most notable change, though, is probably the addition of a feature AMD calls HT Assist. HT Assist is essentially a probe filter intended to reduce the overhead required for the synchronization of cached data across CPUs in multiple sockets. HT Assist reserves space in each processor’s L3 cache, in which it stores an index of where that CPU’s cache lines are being used system-wide. The CPU then becomes “host” of the cache lines stored in its directory. If any CPU needs an update about a particular cache line, it will often know which CPU is the correct host to probe for that information. This reduction in probe traffic can yield big gains in available system bandwidth, as we reported when we saw AMD demo a 4P system whose Stream bandwidth increased from roughly 25GB/s to 42GB/s with the addition of Istanbul processors with HT Assist.”
“Yet Istanbul should be a clear improvement over Shanghai for many workstation-class workloads and most server-class workloads—i.e., those that are essentially parallel and widely multithreaded. The Opteron 2435 manages to deliver this higher performance not just within the same power envelopes, but quite empirically with almost the exact same measured power consumption as the Opteron 2389.”
“This combination yields a nice increase in power efficiency, which was enough to put our Istanbul-based test system in the same territory as our Xeon X5550 system. The competition between the two was remarkably close in SPECpower_ssj, and the Istanbul system required notably less energy to render the Cinema 4D sample scene in Cinebench. So despite that fact that Intel leads in outright performance, the Opteron 2435 is entirely competitive on the power-efficiency front, with lower peak power draw, to boot. Those who evaluate systems strictly on this basis would do well to keep Opterons in the mix.”
“And if you have existing, compatible Socket F servers, the Istanbul Opterons should be an excellent drop-in upgrade. They’re a no-brainer, really, when one considers energy costs and per-socket/per-server software licensing fees.”
ZDNet: “AMD Istanbul: Field Upgrade Only If Your Hands are Nimble”
Unlike Intel’s Nehalem, which requires entirely new and unproven mainboards, systems, chipsets and sockets, the “Istanbul” is based on existing, mature AMD chipsets and uses the existing 1207-pin Socket F, allowing current Quad-Core Opteron systems to be field upgraded rather than requiring entirely new server purchases. . . . . The 24xx and 84xx series CPUs are actually compatible with systems going back four Opteron quad-core chip generations, ever since the 2000 and 8000 series were introduced in August of 2006. . . . . AMD also was able to engineer and manufacture the Istanbul during a period in which it also spun off its entire manufacturing capacity as a separate company, GlobalFoundaries, so the fact they were able to get this product out in record time (six months ahead of schedule, according to the company) to compete with Intel’s Nehalem is a significant achievement.”
Phil Hughes is a senior public relations manager at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.
The Sun Also Rises
Posted by Phil Hughes in 11:44 AM
As part of the build-up toward launching our six-core “Istanbul” processor in June, this is part two of a multi-part blog looking back on the first year of the AMD Opteron™ processor as seen through my personal lens working in AMD Communications.
Starting in third quarter of 2003 following IBM’s launch of its 1U, 2P e325 server, AMD Opteron was slowly building momentum and creating industry buzz. IT customers in the financial services sector who were doing testing on a variety of different AMD Opteron-based systems were seeing impressive results, particularly in 4-processor servers. It was soon clear that the Wall Street IT community talks amongst themselves about technology, and word was spreading quickly about the results they were seeing. The tech press and analysts were starting to catch this feedback indirectly through the Wall Street analysts who heard the chatter from their IT guys.
While this critical end-customer community liked what they were seeing with AMD Opteron, none of them could commit to any large-scale deployments due to lack of enterprise-class systems from a tier-one OEM. The IBM e325 was great for HPC environments, but lacked many of the features required in datacenters.
There are no secrets
This is what I quickly learned about life as a public relations professional in the PC business, it seems like no secret is safe. While not on the same scale, I can sympathize with my PR brethren in the pro sports business trying to keep potential player trades and signings under wraps until negotiations are finalized. Once negotiations between Sun and AMD kicked into high gear the third quarter of 2003, it suddenly became public knowledge. More than two months before Sun announced it planned to offer x86 servers based on AMD Opteron, word was already leaking out. While we were fending off the press with our standard “can’t comment on rumor and speculation” or “you’ll need to contact Sun regarding its future server plans,” there was lots of enthusiasm internally over the upcoming deal with Sun that was code-named “Stinger.”
Even though Sun had little presence with x86 servers at the time, they had a rich history of server innovation and a great reputation for delivering a legitimate enterprise-class solution stack. They were going to launch a diverse portfolio of AMD servers and optimize the Solaris operating system for AMD64. It was a huge morale boost internally and was the start of many activities over the next 6 months that would cast a spotlight on AMD Opteron.
Going back to my earlier point regarding no secrets, two weeks before the announcement, CNET’s Stephen Shankland broke more news on the Sun-AMD partnership, including the announcement date. After that point, the only unknown was how many and what form factor servers they were planning to offer. At his keynote at Comdex, CEO Scott McNealy revealed Sun’s plans for a broad partnership with AMD, details of that announcement can be found here.
A strong finish
While Sun would not begin shipping AMD Opteron servers until early the next year, the announcement gave us more of what we really needed at the time: credibility. It was a turning point in demonstrating that AMD Opteron was more than just an HPC solution.
As 2003 came to a close, AMD Opteron had been in the market for about eight months and market share numbers were starting to show some impact. While most of the early AMD Opteron servers were shipped into HPC environments, there was enough traction for analyst firms such as IDC to start tracking systems that were based on AMD Opteron processors. In December 2003 IDC issued its third quarter system share numbers and a number of reporters latched on to the fact that the AMD Opteron processor was shipped in more servers than Intel’s Itanium processor.
Our intent was never to go after Itanium’s market share as we knew we couldn’t be successful if that was our goal. Both Itanium and AMD Opteron were 64-bit processors but that’s where the similarities ended. AMD Opteron was perhaps the best x86 server processor available by the end of 2003 and our objective was to bring 64-bit into the mainstream server market, not the proprietary space Itanium was playing in. Intel soon recognized this and my next blog will discuss the new battleground that emerged in 2004.
I’ll conclude with this excerpt from an end of year wrap-up story InfoWorld ran at the end of 2003:
“AMD’s Opteron processor truly is the Little Engine That Could. When Opteron launched in April, the Intel x86-compatible 64-bit processor faced a steep uphill battle against Intel’s Itanium. By year’s end, IBM and Microsoft delivered on early promises, and Sun joined the bandwagon. Will Dell and HP be next, or can’t they see Opteron from where they sit behind Intel?”
Up next: “Swordfish anyone?”
Phil Hughes is a senior public relations manager at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.
AMD-on-AMD action
Posted by Phil Hughes in 6:31 PM
A couple of weeks ago at a press event in Sunnyvale celebrating the sixth anniversary of the AMD Opteron processor launch in 2003, we provided details on Direct Connect Architecture 2.0 and our 2010 server platforms code-named “Maranello” and “San Marino”.
Both “Maranello” and “San Marino” will be “AMD on AMD” platforms, meaning they will leverage chipsets developed by AMD. Currently shipping AMD server platforms use chipsets from Nvidia and Broadcom. Both Nvidia and Broadcom will continue to support Socket F platforms through 2010, including upcoming systems based on our six-core “Istanbul” processors.
When asked at the press event about our relationship with Nvidia and Broadcom in the context of our server platforms, Vice President and General Manager of AMD Server Business Unit Pat Patla stated the following: “For 2010, moving forward, the solutions coming out from AMD will be AMD and on AMD at this time. We don’t expect to see new chipsets from NVIDIA or Broadcom for server implementations in 2010.”
Unfortunately, this led to some speculation that there were licensing issues between AMD and Nvidia. At no time did AMD offer any comment related to future server processor licensing agreements with Nvidia or any third-party chipset provider.
Dating to the ATI acquisition in 2006, the long-term AMD plan remains consolidating on our own chipsets for future server platforms. As such, it was a mutual business decision between AMD and our third-party server chipset providers that they would not develop chipsets for future AMD server platforms. This was a decision specific to the server business unit.
We are very excited about the opportunity to bring new platform-level innovations to the x86 server market later this year, starting with our “Fiorano” and “Kroner” platforms in the second half. Some of those new features include:
- IOMMU – Increased performance through enhanced I/O and improved security through hardware isolation of VMs
- HyperTransport 3 – Increased platform-level I/O bandwidth
- PCI Express Gen 2 – Optimized for Accelerated and Stream computing environments
It started with a single core
Posted by Phil Hughes in 1:01 PM
On April 22 we celebrated the sixth anniversary of the AMD OpteronTM processor launch, the server processor that showed just how ripe for innovation the x86 server market was early this decade. As we move closer to the launch of our six-core “Istanbul” processor in June, below is the first post of a multi-part blog that provides some perspective on the first year of AMD Opteron through the eyes of someone intimately involved in driving awareness for the processor since its launch.
"Hammer" time
I started with the AMD PR team in October 2002, about 7 months before AMD Opteron, then code-named “SledgeHammer,” officially launched at the Hudson Theater in New York City on April 22, 2003. Things such as energy-efficiency, multi-core, performance-per-watt and cloud computing were not part of the discussion back then.
The bridge to 64-bit computing … and credibility
Our lead message then: TCO and ease of migration relative to the fact that we could offer 64-bit computing on an industry-standard architecture with backwards compatibility with existing 32-bit applications. It was a simple one for our audiences to appreciate considering our competitor was going the path of Itanium to get to 64-bit computing. How simple? Well, our tagline initially was “Simplifying Business” and the supporting points were all about investment protection, flexibility, performance and industry support. That last one is probably the most important one for me to expand on a bit further.
The biggest challenge for AMD in the server business in 2003 was credibility. Lots of IT guys loved our desktop processors but we didn’t have any real history in the enterprise. We had only been offering server processors since 2001 (AMD AthlonTM MP processor) and none of our system support for that processor was from tier one OEM customers. In fact, in most of the pre-launch communications we did, our biggest credibility proof point was our relationship with Microsoft and other ISVs. Even though most of our early adopters would ultimately be Linux guys, Microsoft’s support provided credibility to a server solution that did not have tier one OEM support until the day of our launch.
The “bridge to 64-bit computing” image we frequently used in press meetings prior to the 2003 launch
All we need is one. Right?
From a PR perspective, the Bay Area press and analyst tour Kevin Knox, then AMD’s director of commercial business development, and I conducted in February 2003 included several meetings with a strikingly similar dialogue. Nearly all recognized and agreed that the competition made the wrong move with its initial stab at addressing 64-bit computing, but most could not accept the idea of AMD Opteron being successful until tier-one OEMs came on board.
More frustrating at that time was the fact that we were getting seed systems into end-user hands, but no high-profile end customers would go and talk about their positive experience with AMD. Many reporters indicated to us that getting a couple of CIOs to talk on our behalf would generate instant credibility, but unfortunately without tier-one support, their credibility could also be at risk by endorsing AMD Opteron.
When we did launch AMD Opteron, IBM joined us on stage in NYC (you can view the opening launch video here) and said they would be offering a 2 processor system targeted at the HPC market later that summer. While not exactly what we were hoping for, it was definitely a shot in the arm. We figured once one OEM came on board, others would follow suit. Unfortunately that was not the case. We immediately became seen as primarily a solution for HPC and research. No one felt we had the chops to do mission critical, because you couldn’t find AMD Opteron in a mainstream, enterprise-class server from Dell, HP, IBM or Sun. Also, one of our lead messages around 64-bit computing really wasn’t pulling through well because there just weren’t a lot of applications at that time that could take advantage of it.
Buzzkill
Still remains one of our favorites on the PR team
Here we were in the summer of 2003, just a few months removed from the launch and we’re approaching LinuxWorld, which was an emerging enterprise tradeshow at that time. We had no server market share to speak of, and a single OEM SKU that had yet to launch.
My challenges usually came when I would get blindsided by a reporter who had recently spoken to and OEM or system builder who said customers were not asking for AMD Opteron and while they had looked at it, they had no plans to support it. As a PR guy, I’m glad we had IBM as a proof point but that one was starting to wear thin and it kept feeling like we were waiting for the other shoe to drop with the press coverage.
COMING NEXT: The Sun also rises
Phil Hughes is a senior public relations manager at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such linked sites and no endorsement is implied.