Well folks, here are the final set of answers.  If you haven’t noticed, with each of the 3 previous rounds we had tons of comments come in as questions, so it is probably worth a read through of the comments because there might be some other nuggets of information for you.

Part 1 Answers

Part 2 Answers

Part 3 Answers

“Hi AMD-people, when scaling from 12 to 16 cores (MC to Interlagos) will you be able to still maintain the thermal design envelope without cutting down on clockspeeds? Sorry for my bad english ” – Kenny

Well, as we have said before, we are not revealing some things until launch (clock speeds, cache sizes, pricing, benchmarks or the launch date) but I can say that from a clock speed perspective the base clocks speeds will be very competitive in our opinion.  Please keep in mind that we will have AMD Turbo CORE technology in these processors, so the actual frequencies can be higher than the base frequencies in many, if not all, workloads.  We have made some improvements to Turbo CORE technology that we cannot share just yet, but I can say that the way we have implemented it with “Bulldozer” overcomes some of the limitations in previous turbo implementations. 

TDPs are all planned to be the same between current AMD Opteron™ 6100 and 4100 series processors with the new “Bulldozer”-based processors (codenamed “Interlagos” and “Valencia,” Respectively).

“How much resemblance does your today’s Bulldozer architecture have to the original design?” – Tye

As you are aware, when we initially designed “Bulldozer,” we were working with a more modular processor design.  The original “Bulldozer” design was 45nm.  But as development progressed, it became clear that the 45nm design that we had been working on was not going to be as competitive as we would have liked.

With the world watching your every move, all product decisions become very public (how many articles have you seen about “Bulldozer” in the press?) It is never an easy decision to make substantial changes to your product roadmap, but sometimes you have to make the tough call because it is the right thing to do. It helps when your partners are behind you and agree with the changes – sometimes a little short-term pain is necessary for a long-term gain.

We obviously aren’t going to get into specific design changes, but we think that the 32nm “Bulldozer” can bring a lot of benefit to our product due to smaller transistor size (which can help drive down the power envelope).  By going for lower power, we hope to give you more room for compute cores, FP capabilities and more.

In addition, bringing the AMD Opteron™ 6100 Series processors to market allowed us to deliver an MCM design to market, which validated that technology.  Having the SR5600 series chipsets, with more than a year under their belt, means that platform validation should be much easier for our partners. One of the biggest beneficiaries of a 32nm design is not necessarily a new technology added to the processor as much as it is the socket, chipset and platform work that happened ahead of the “Bulldozer” release.

“How the cache subsystem looks like and how much L2 and L3 cache memory Bulldozer have?” -  Marcin Szymanowicz

As I stated in the first answer, we won’t give cache sizes, but I can talk about how the caches are laid out with respect to the module.  The L1 cache is comprised of two components, an instruction cache and a data cache.  The instruction cache is 64K and is shared between the 2 integer cores in the module.  The data cache is 16K and there is one dedicated data cache for each core in the module.  The L2 cache holds a mixture of both data and instructions and is shared by the two integer cores.  If only one thread is active, it will have access to the entire L2 cache.  The L3 cache is shared at the die level, so on the “Interlagos” processor you will have 2 separate L3 caches, one for each die.  On the “Valencia” processor you will have one L3 cache because there is only one processor die.

“Why don’t you for example publish the list of patents related to technologies implemented by Bulldozer, and the list of patents which appeared during the last years but are NOT implemented by Bulldozer? In addition, a simple list of non-patented technologies implemented by Bulldozer would be very helpful as well.” – Square Symbol

If only it were that easy.  AMD has hundreds of patents each year based on the research of our engineering teams.  There are people who spend many hours combing through public records trying to piece together all of the details of what our architecture might or might not be.  For competitive reasons, we aren’t going to publish such a list, before or even after a processor has been released. For companies like AMD that live in a highly technical world, many of these details help give us a competitive edge.  To customers, it’s not usually the technology in the processor, but it’s those “3 P’s” that matter: What is the Price? What is the Performance? And what is the Power efficiency? 

“As far as power usage goes, from what I understand BD is supposed to be taking power management features to a level of granularity that hasn’t been seen yet with consumer/business grade CPUs. Will those new features be available to current MC users or will a platform upgrade be necessary?” – Jeremy Stewart

”Bulldozer’s” features that are designed to increase power savings reside in the processor and BIOS, not the platform. Many customers with existing AMD Opteron™ 6000 Series platforms should be able to upgrade to “Bulldozer”-based “Interlagos” processors when they are available.  The reason that I say “many” and “should” (in addition to having lawyers review my blogs) is that an upgrade is not necessarily something that AMD can guarantee – your ability to upgrade will depend upon your individual system configuration, and that’s something we cannot control. 

“Interlagos” processors will be socket compatible and are designed to have the same power/thermals, so it should work with existing thermal solutions.  However, we don’t control the system BIOS, nor the system support.  It is best for you to check with your system or motherboard vendor to determine whether your system can be upgraded. 

The vast majority of servers are not upgraded, but the more important question is whether the platform changes between the current processors and the new ones.  Having specified the AMD Opteron™ 6000 series platform to work with both processors, there should be no need for the platform to change for many systems, making it far easier for customers to manage multiple generations of processors side by side.

John Fruehe is the Director of Product Marketing for Server/Workstation products 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. This blog contains forward-looking statements.  Forward-looking statements are generally preceded by words such as “plans,” “expects,” “believes,” “anticipates” or “intends.”  AMD Investors are cautioned that all forward-looking statements in this blog involve risks and uncertainties that could cause actual results to differ materially from current expectations.