AMD's Barcelona: Why we haven't published benchmarks

A month ago we were able to present to you a rare look at AMD's forthcoming roadmap, detailing everything from new plans for the mobile space to giving a better idea of AMD's reasoning behind the ATI acquisition. We left that article with a positive and hopeful note for AMD:

"For a while we had lost confidence in AMD, like many of you had as well, and although AMD's position in the market hasn't changed we are more confident now that it can actually bounce back from this. Intel seemed to have the perfect roadmap with Conroe, Penryn and Nehalem all lined up back to back, and we saw little room for AMD to compete. Now, coming away from these meetings, we do believe that AMD may have a fighting chance. Over the coming months you'll begin to see why; it won't be an easy battle, but it will be one that will be fought with more than just price."

A strong roadmap alone does not make for a successful company; we need to see near term execution as well. For AMD, that means Barcelona has to be competitive. The interesting part of AMD's disclosures as of late is that as much information as AMD has given us about its roadmap for 2008, 2009 and beyond, we have little to no details about when we can expect Barcelona and how fast it will be.


When we headed out to Taiwan, a country of leaked processors and benchmarking opportunities, for Computex we surely expected to return with some Barcelona performance figures. We were hoping we'd come back with the very data that AMD hadn't allowed us to get ourselves when we visited the company over a month ago. And while there were some performance results reported from Taiwan, there was an eerie silence about AMD's updated micro-architecture.

We were determined not to leave the island without running at least one test on Barcelona. We worked long and hard, and we were finally able to spend some time alone with Barcelona in Taiwan. But the story doesn't end there; it's unfortunately not that simple.

Motherboard Problems

We know that Barcelona works and runs benchmarks, as we saw back at AMD in May. But the demos that AMD ran were on its own motherboards, not on motherboards from its partners. AMD's partners just recently received their first "production quality" Barcelona samples, and as expected, the current boards required some heavy BIOS work before the new chips would even work, much less perform up to the expectations set by AMD.

The motherboard we tested on had minimal HT functionality and wouldn't run at memory speeds faster than DDR2-667; most 3D video cards wouldn't even work in the motherboard. Memory performance was just atrocious on the system, but the motherboard manufacturers we worked with attributed this to BIOS tuning issues that should be fixed in the very near future.

In the end, performance was absolutely terrible. We're beginning to understand why AMD didn't let us test Barcelona last month. It's not that AMD is waiting to surprise Intel; it's that the platform just isn't ready for production yet.

More Barcelona
POST A COMMENT

64 Comments

View All Comments

  • JackPack - Monday, June 11, 2007 - link

    Employee morale at AMD must be pretty much low at this point.

    Every time an AMD spokesperson has to say "Barcelona is on track for a summer launch" or "regaining the performance lead this year" when they don't believe it themselves, it takes a toll on the employees.

    It reminds me of what one AMD spokesperson said in July 2006, right after Intel's Woodcrest launch: "We remain supremely confident in both our product architecture and road map, which remains unchanged."

    Why don't they just come clean about it?
    Reply
  • Roy2001 - Monday, June 11, 2007 - link

    This reminds me the Iraq officer who claimed Iraq troops were defeating US troops when US army's tanks entered Bagdad. Reply
  • OddTSi - Monday, June 11, 2007 - link

    Maybe AMD should hire good ol' Baghdad Bob. At least he'll give us all a few good laughs. Reply
  • shabby - Monday, June 11, 2007 - link

    Dear stock holders, our product sucks and will probably be delayed.

    Do you really expect a company to say this?
    Reply
  • TA152H - Monday, June 11, 2007 - link

    Barcelona having problems is so predictable it was almost a certainty. AMD doesn't have a great reputation for manufacturing, and currently their fastest parts are still 90nm, not the 65 nm lithography the Barcelona is made on. Since their well understood Athlon 64 can't reach high clock speeds on this process, naturally a newer design that is not well understood will be even worse. On top of this, AMD made a typical marketing mistake and made the Barcelona a native quad-core, something even Intel with much better manufacturing ability thought was currently too difficult. So, Barcelona running at 1.8 GHz is not terribly surprising considering everything. It doesn't mean, at all, that a dual core would be suffering from that clock speed either. There was a reason why Intel wouldn't do it yet, but of course AMD knew better. Hector has to go!

    I don't know why anyone at AMD would panic, except for the fact they have realized Intel was right about not rushing to native quad core. On the plus side, if they created a new socket, a dual-core version of the Barcelona would be a much better 8-socket than the Opteron, which was horrible after four sockets. So, if they have to eat crow and release slow quad-cores, at least the 8-way performance will be really good and might save them.

    The performance of this processor is something else no one needs to worry about. I still think they are making a mistake splitting the integer and floating point clusters; it's a waste of resources and not many people care about x87 anyway. But the Barcelona has non-speculative memory reordering, whereas the K8 was incredibly primitive and memory access were all issued in-order. It's not up to the Core 2 standard, but it's equivalent to the Pentium Pro - Pentium III now in that regard, and that alone should make a nice improvement. With all the other improvements involved, I don't think anyone needs to worry about the IPC much.

    The only thing I would worry about is the native quad core stuff. I don't know if AMD will tackle that easily, because Intel decided it was too difficult. But, for the dual core desktop processors, I can't imagine any serious long term problems. Maybe a delay while they find the critical paths and improve their manufacturing technology, but panicking just seems extreme to me. The IPC will surely not be anything anyone needs to worry about. It may not be quite as good as a Core 2 in certain workloads, but it will certainly be a great improvement on the K8 and will narrow the gap in the worst case, and possibly pass it in others. Certainly as you add sockets, it should do considerably better. And even on single sockets, you'll have a great advantage in IGP solutions since the processor can now sleep and use a lot less power.

    I wouldn't worry too much if I were an AMD stockholder. I don't think there's anything too surprising, and nothing insurmountable, although native quad-core might not be feasible in the relatively short term and might compare unfavorably to Intel's solution. But, it's not a huge market yet anyway.
    Reply
  • LTG - Monday, June 11, 2007 - link

    quote:

    quad-core...might compare unfavorably to Intel's solution. But, it's not a huge market yet anyway.


    Are you kidding? There is a huge market for quad core servers and it's where the profit is.

    AMD will remain under huge price pressure until 08 if this article is even close to true.

    Not good, not good at all.
    Reply
  • neweggster - Wednesday, June 13, 2007 - link

    No way is Quad core server market as strong as you think. The problem here is that many many server specialists seek out alot more then just small real world performance upgrades. The research and product bids go towards many factors such as power management, performance equations to stability and so on. Meaning Quad core server market will hardly be pulled on hard enough till performance and other things are well enough to make such huge costly changes. Server people tend to stay with what works and for years at a time, wasting money on new products that aren't proven to offer huge benefits just yet makes it a worthless business transaction. Reply
  • TA152H - Monday, June 11, 2007 - link

    Actually, there isn't. There haven't even been quad cores out for very long, and it takes a while for server folks to move over to new technology like this is. The big advantage AMD has is they don't even need quad cores, and they are better off with 4 x duals instead of 2 x Qs. Intel with their FSB scale much more poorly to four sockets, and it wouldn't buy anything. AMD would presumably have a better solution with four sockets and dual cores than two sockets and quad cores since you'd increase memory bandwidth, unlike Intel's solution.

    Naturally, Quad by four sockets would be better, but that's comparing AMD with AMD. I'm sure they'll release something for that anyway, and if it runs a little slower in clock speed until they fix things, it's not a big deal because Intel really has nothing in that market worth a damn. AMD may run at slower clock speeds, but Intel's entire system architecture is horrid for it. So, I wouldn't call the 16 processor area a disadvantage for AMD at all. As I said, the eight processor arena should also be fine for AMD, just it will probably have four sockets instead of Intel's two. Again, their system architecture gives them a big advantage in having eight cores on four sockets, whereas Intel's doesn't so it doesn't make much sense for them. Case in point, AMD's 4 x 4 or whatever stupid name they have.

    Also keep in mid that people that buy servers like to think ahead, and in a general sense you'd prefer your servers to be as similar as possible. I really like Intel solutions because their chipsets and motherboards are excellent quality, and I don't like to buy AMD anymore because I hate NVIDIA chipsets, and I don't trust ATI yet (they always seem late, slow and lacking features), but AMD's system architecture is so much better for four socket solutions, I can't see buying Intel now or in the near future. A lot of people buying servers, or planning to (as is most often the case) may look at see the Barcelona is a little slow right now, but Intel's solution is slow for quite some time; AMD should solve their clock speed problem well before Intel solves their system architecture problem.

    I don't think this is a big deal. I think it's a big deal only in the sense that it seems that nothing AMD can do right now works, and everything gets blown out of proportion because it's always gloom and doom with this company. It's a real concern, of course, when a company has nothing but bad news, but I don't think this particular problem is really huge like it might initially seem. The constant problems though, when added up, surely do present a pretty bleak picture. But it wasn't that long ago Intel was making mistake after mistake. The difference is, everyone knew Intel would come out of it because they have so much talent and ability, but AMD is much smaller, so it's a bit scarier. But, remember, they have IBM on their side with manufacturing, and that should help.
    Reply
  • CyberHawk - Monday, June 11, 2007 - link

    True!

    But, with all said, and all the unofficial benchmarks out there I also expected some real numbers. Looks like we'll have to wait a little longer.
    With AMD pricing right now we do not have to worry. I hope that it turns out well, it is not a no-go situation.
    Reply
  • Gary Key - Monday, June 11, 2007 - link

    quote:

    But, with all said, and all the unofficial benchmarks out there I also expected some real numbers. Looks like we'll have to wait a little longer.


    We tried/completed close to 30 benchmarks on an early 2P system but the results just were not good but this was based on a very early BIOS release. While we wanted to print real benchmarks, the numbers would not have meant anything considering the early BIOS issues. Hopefully, over the course of the next couple of weeks we should see tremendous progress being made in the BIOS tuning efforts.

    The limited demos that AMD's partners showed during Computex were on a 1P system also, we had the same results with two benchmarks that ran without issue but felt like since Barcelona is destined for the server market that 2P results would have been more meaningful. What was more telling than our early benchmarks was just the fact that the platform is not ready yet. A fact that really floored us considering all of the press information about AMD being on schedule. While we feel confident that AMD could meet a limited (as in very limited) late July roll out of Barcelona, until core speeds improve, we do not see this as being beneficial except from a PR viewpoint. However, maybe that is what AMD really needs at this point until volume shipments are occurring. :)
    Reply

Log in

Don't have an account? Sign up now