AMD's Barcelona to Outpace Intel by 50%

Gr8Apes writes "AMD is upping the performance numbers for Barcelona by stating that "Barcelona will have a 50% advantage over Clovertown in floating point applications and 20% in integer performance 'over the competition's highest-performing quad-core processor at the same frequency'". AMD also claims that the new 3.0 GHz Opterons beat comparable Intel Xeon 5100 series processors in three server-specific benchmarks (SPECint_rate_2006, SPECint_rate2006, SPECompM2001) by up to 24%."

Comments

Where would you want to live?

[Anonymous Coward]

Barcelona or Clovertown?

Nice attempt, AMD.

[EveryNickIsTaken]

Barcelona will outpace Intel's "current-gen" chips by 50%, not the ones that are currently in production. Nice attempt by AMD to become relevant again, though.

Re:Nice attempt, AMD.

[Anonymous Coward]

Je demande un recompte!

Re:Nice attempt, AMD.

[minginqunt]

Well, quite.

But you know what they say about lies, damn lies, and benchmarketing.

Or, if they don't say it, they ruddy well should.

Re:Nice attempt, AMD.

[Chris Burke]

Yeah, and notice that they say "at the same frequency", when Intel currently has a frequency advantage (not as big as P4, but then again Core 2 isn't an IPC dog like P4 was). Not that I expect any minor improvements Intel makes in the next 60 months to produce their own 50% leap in performance, this comparison still seems very suspect. As in pure marketing BS.

However AMD doesn't need to attempt become relevent again. They are currently very relevent. Did Intel become irrelevent when they were behind AMD on performance? No. In the past, AMD did lose more by not having the performance crown, and one could certainly imagine the momentum they were gaining in the K7 days fading quickly if Intel had come out with a superior chip. But today, AMD has both the marketshare and the OEM support to be merely competitive performance-wise and still be relevent. So they lose out at the top speed grades. If they can continue to match up their products to Intel's at lower speed grades, and they will, then they will continue to be a good choice for many people, and will definitely still be relevent.

Re:Nice attempt, AMD.

[Chris Burke]

Obviously that was supposed to say "in the next 6 months", no "60". I'm betting Intel can produce at least a 50% speed improvement in 5 years. :P

Re:Nice attempt, AMD.

[MrFlibbs]

Note that AMD's claim is to be faster "at the same clock". When the P4 was pushing clock speeds into oblivion, AMD stressed the point that clock speed is irrelevant -- what really counts is how fast the system runs your software. How you get there is quite beside the point. How odd that AMD is now using clock speed as a key indicator.

Intel is already shipping 3GHz Clovertowns, and the article states that AMD has not released the Barcelona clock targets. It they ship substantially below 3GHz (2.4?), then Barcelona will probably still win on FP benchmarks (barely) but lose on everything else. This suggests it will be more competitive, but not compelling.

Re:Nice attempt, AMD.

[Skadet]

Barcelona will outpace Intel's "current-gen" chips by 50%, not the ones that are currently in production.

Could you explain how current generation chips differ from those that are being produced and shipped today?

Re:Nice attempt, AMD.

[Gr8Apes]

Note that AMD has 3GHz Opterons out now. The frequency advantage at the moment is slim to non-existant in shipped CPUs.

In any case, what matters is what Barcelona will ship at, which has not yet been specified. In any case, if Barcelona lives up to AMD's stated expectations on performance, it will be a killer CPU. Your statement about Intel's potential improvement leaps are spot on, and fall into Inforworld's Tom Yager's statements about Intel which are essentially phrased as "Core 2 is Intel's last hurrah". Why? Because Intel is essentially running on 10 year old technology and is rushing to catch up, despite some of the nifty architectural things they did recently to speed up C2D (integrated L2 cache for example).

I also believe that Intel is now following AMD's lead by leaving extra headroom for those that prefer to OC their CPUs and concentrating more on TDP and stability. I've noticed that Intel's chips since P4 are certainly more stable, while my rather severely OC'd AMD CPU occassionally (twice this year) shuts down, most likely due to heat or a RAM instability (since the shutdowns happen during low usage periods at night, I'll bet the 20% OC'd RAM is probably to blame).

Basically, right now Intel owns the crown, but they own it while comparing to AMD's last gen CPUs which are 3+ years old.

Re:Nice attempt, AMD.

[Yoooder]

Does anyone here actually expect either IBM or AMD to win this tug of war any time soon? I'll keep buying the best proc for the money, for the past few years it's been AMD, now it's Intel. It'll change again, and again after that...

Re:Nice attempt, AMD.

[Gr8Apes]

And then add in the power savings, among other things. I actually expect Barcelona to have an even larger gain on multi-threaded applications, it's single-threaded apps that will have the smallest gain or be roughly even.

I'm seriously wanting to see what video editing and rendering performance numbers will be between the two. Clovertowns will have inherent cache flushes, as there are 2 independent caches each shared by 2 cores that need to stay in sync. Barcelona has a shared L3 cache among all 4 cores and independent L2 caches.

Re:Nice attempt, AMD.

[donglekey]

Intel is also leaving a big gap for AMD to fill with the price of their quad core cpus. A grand? AMD should be able to compete with that. I just want a shuttle with four cores, and if AMD delivers with procs that are 'only' $500 that will be pretty significant.

Re:Nice attempt, AMD.

[egomaniac]

Note that AMD's claim is to be faster "at the same clock". When the P4 was pushing clock speeds into oblivion, AMD stressed the point that clock speed is irrelevant -- what really counts is how fast the system runs your software. How you get there is quite beside the point. How odd that AMD is now using clock speed as a key indicator.

Ummm... they're not. If they were using clock speed as a metric, they would be saying "Look! We're running at 3.5GHz and Intel is only running at 3GHz!" while completely ignoring the actual performance -- exactly what Intel did all those years. They are instead talking about performance-per-clock-cycle, which (according to this) means that a 2.66GHz AMD chip would be considerably faster than a 3GHz Intel chip. We can expect them to continue touting the overall performance rather than raw clock speed, since they look better from a performance standpoint and worse from a raw clock speed standpoint.

How is that different than what they've been saying all along?

Re:Nice attempt, AMD.

[fullmetal55]

another thing, yeah they don't need to become relevant again, they are still. But, The "At the same frequency" has me a bit confused. since AMD has often had a higher performance per frequency, That's not news, however the "same frequency" did not = "same pricepoint" and a similar freqency product from AMD cost more than from intel. example a 2 Ghz Athlon XP (3000+) was more than a 2.0 GHz P4. while the similar pricepoint 2000+ ran at 1.66 Ghz... I'd like to see a price-point comparison...

Re:Nice attempt, AMD.

[Overly Critical Guy]

More like an attempt by AMD fanboys to make up for the past two negative articles on AMD by declaring that they will "Outpace Intel by 50%" in a Slashdot headline. Just to make themselves feel better about their choice of processor team to root for. Me? I go for whoever's best, and right now that's Intel.

Re:Nice attempt, AMD.

[logicnazi]

Actually if Intel is running on 10 year old chip design technology (I presume this is primarily a remark about the FSB) then this suggests they have the potential to *radically* increase performance over AMD. If Intel's process advantages and chip design teams can actually gain a performance advantage while using 10 year old technology they can sit back and pick the low hanging fruit (changing to modern methods) and gain huge performance boosts while AMD has to do truly innovative things to gain any performance increases.

Frankly this sounds more like fanboi talk than a serious analysis. If your goal is to diss Intel and give AMD props then saying they are using 10 year old technology makes sense. If you are actually trying to argue that AMD's future is much brighter than Intel's it's totally non-sensical. If Intel can gain huge performance benefits just copying stuff AMD is doing now while AMD has to make huge advances just to stay competitive I know who I would put my money on.

Re:Nice attempt, AMD.

[Chandon Seldon]

AMD's release will be first, so they'll be in the lead for a little while - in exactly the same way that Intel's Core 2 Duo release came before Barcelona.

In the mid 90's, Intel stuff was pretty obviously better. Then AMD was pretty obviously better for a while. Now it's really interesting - they're basicly neck and neck with each pulling ahead with new releases and then falling behind again.

Re:Nice attempt, AMD.

[didde]

Frankly this sounds more like fanboi talk than a serious analysis. If your goal is to diss Intel and give AMD props then saying they are using 10 year old technology makes sense. If you are actually trying to argue that AMD's future is much brighter than Intel's it's totally non-sensical. If Intel can gain huge performance benefits just copying stuff AMD is doing now while AMD has to make huge advances just to stay competitive I know who I would put my money on.

Err, you're assuming Intel in fact has a brand spanking new architecture up its sleeve, ready to be rolled out at any given moment. State your sources please. Otherwise, let's not speculate on the future x 2.

Re:Nice attempt, AMD.

[644bd346996]

Intel has been doing a lot of work to get around the relative limitation of not having an on-board memory controller. When they integrate the memory controller, quite a bit of that work will be irrelevant. Sure, being smart about how you cache things is good. But intel has done what they can to reduce their dependence on a bottleneck. When that bottleneck is removed, they won't necessarily be able to scale to full use of the new bandwidth.

I think it is AMD that has been increasing the performance by making drastic changes (64 bit, on-die memory controller, HT bus) whereas intel has been using their fab capacity to their advantage, by including big caches (for x86) and being the first to a die shrink.

Re:Nice attempt, AMD.

[Gr8Apes]

It's not purely about FSB, although that is a major short coming for multi-core designs. Intel won't radically improve their performance by going with a hypertransport design, but they'll improve their scalability (the quads are already beyond the maximum FSB - see the Xeon benchmarks for Mac Pros that's available). The 10 year old design refers to their falling back to the P-III core and then adding in a bunch of cache logic, fab processes and other more minor tweaks to improve that core's performance to what it is today in the C2D. The next major improvement for Intel's performance scalability is with Nehamal (I think it is) when they're supposed to finally implement a HTT type solution for memory communication, although I was unclear whether it also addresses multi-CPU/core communication. I would think it would, as anything less would be moronic.

AMD has other items up its sleeve. Not all are pure performance improvements, but interesting things like hardware virtualization. This will become more improtant in the future, and may even filter down into the consumer space. Wouldn't it be great for every app to essentially have its own virtual world? That would indicate a much lower potential for harm inherent to the OS architecture. (Note, odds are MS won't go this way because it breaks forced dependencies on the OS, but it certainly opens up new worlds for others)

Intel can't copy AMD's advances. They'd have to license some tech to do so, and that's not very likely (I believe HTT for instance belongs to AMD). Intel's current big performance leap occurred because they dumped P4 and followed AMD's approach. In any case, Intel did have a lot of smart people, but with the 10,000 personnel layoffs, how many of those smart peopel won't look for better jobs elsewhere? (Layoffs don't exactly inspire confidence for a company's future) Of course, AMD announced they may have to do the same after Q1's results.

Re:Nice attempt, AMD.

[Gr8Apes]

Intel has already announced major price cuts for the Fall. I can't see how this does anything other than hurt Intel and current sales. For instance, I'm sure I'm not the only one that sees that and goes... hmm, I might have considered a new Intel CPU, but the prices are dropping significantly in a couple of months. And then there's the AMD roll out at the same time, maybe I'll wait and see what that brings to the table.

It's going to be very very interesting.

Re:Nice attempt, AMD.

[Chris Burke]

How is that different than what they've been saying all along?

Because they've been saying all along that it is overall performance that matters, while here they are keeping the discussion only in the realm of IPC. Performance in Instructions/second = IPC * Hz. If you talk only about IPC or only about frequency, then you're leaving out half the equation that results in the real number you want.

So while they are at least talking about benchmark scores and actual performance (can't really talk about IPC without them), the problem is that they are leaving out the effect of frequency. In the absence of any other information, "40% more IPC" isn't really any more informative than "40% more frequency". We can guess what frequencies Barcelona might come out at, and thus make some use of this information, but it is definitely a difference from AMD's old statements.

Re:Nice attempt, AMD.

[Anonymous Coward]

Actually, Intel has put out quite a bit of public information about their "Nehalem" core. This architecture adds in one of AMD's biggest advantages - the on-die memory controller. (Also, an L3 cache) This is set to debut in 2008 on a 45nm process. I sure hope AMD has some tricks up their sleeves because I like the competition, but Intel certainly is making some phenomenal progress.

    (I'd cite sources, but I'm lazy... search for Nehalem and you'll see.)

    Anyways, this 'newer, friendlier' Intel has said publicly that this is the strategy they pursue - there is a two year cycle for new architectures which is staggered with a two-year cycle for new fabrication facilities. And 2008 is shaping up to be an impressive one for them from all of their press. .. Then again, it is for now just press. :)

Re:Nice attempt, AMD.

[zephyrprime]

>despite some of the nifty architectural things they did recently to speed up C2D (integrated L2 cache for example). The L2 has been integrated since the P3 brainiac.

Re:Nice attempt, AMD.

[Targon]

I believe that AMD is saying how Barcelona will perform compared to the chips available today, not the chips Intel will have down the road. This is their way of saying that if Intel suddenly is able to release much faster chips between now and Barcelona's release, this claim by AMD won't reflect THAT.

So, current Core processor machines are what AMD is making claims about, not the 45nm chips that Intel is currently working on but has not released yet.

Re:Nice attempt, AMD.

[Endo13]

Err, you're assuming Intel in fact has a brand spanking new architecture up its sleeve, ready to be rolled out at any given moment. State your sources please. Otherwise, let's not speculate on the future x 2.

Exactly. Netburst was their new technology that was supposed to be "the future" and we all know how well that worked out. It really is quite impressive that they've done as well as they have at evolving the P3 architecture. But as an AMD fan, this scares me because I don't know what they're going to do next if/when AMD manages to blow them out of the water again like they did back in the early P4 days (and right now I'd say it looks imminent). Despite being an AMD fan, I like Intel as well, and I don't want to see them die in the next 10 years. But it scares me because they've already played their trump card, and I'm not so sure they have another. Here's to hoping that AMD tops Intel again, but not by so much that Intel starts losing lots of market-share again.

Re:Nice attempt, AMD.

[Anonymous Coward]

Netburst was their new technology that was supposed to be "the future" and we all know how well that worked out

Actually, I thought "VLIW" & Itanium was supposed to be the future, and P4 was a product of the backup team. Core was a result of the laptop group that knew they couldn't make the P4 low-power so they cobbled something together using the best of P4 & P3 technology that wound up spanking the big brother in the real world.

Those two big misses let AMD catch up and surpass Intel (not to mention RDRAM & other smaller disasters born of Intel hubris).

Re:Nice attempt, AMD.

[PitaBred]

They probably mean that they'll be running 50% faster than the chips that Intel will have out at the same time (ie, the Penryn cores I believe). At least, that's my take on it.

Re:Nice attempt, AMD.

[Gr8Apes]

>despite some of the nifty architectural things they did recently to speed up C2D (integrated L2 cache for example).

The L2 has been integrated since the P3 brainiac. First, you're correct. Integrated L2 cache across 2 cores with cache lookup and control optimizations. Next.

Re:Nice attempt, AMD.

[Gr8Apes]

Exactly. In 2008, Intel will finally address some of the biggest performance features in AMD's architectural stable. Whether it will be as effective as AMD's solution will remain unknown until the first CPUs come out and can be benchmarked.

Amazingly enough, 2008 is when AMD's 45nm process should hit full stride (although AMD historically has had some issues with die-shrink processes). AMD also still has SOI, which Intel won't be able to counter until that just announced new process (forgotten the name) hits sometime in late 2009 or 2010. And I'm sure AMD will be sitting still architecturally and technically while Intel catches up.

Re:Nice attempt, AMD.

[Jerry Coffin]

So while they are at least talking about benchmark scores and actual performance (can't really talk about IPC without them), the problem is that they are leaving out the effect of frequency. In the absence of any other information, "40% more IPC" isn't really any more informative than "40% more frequency". We can guess what frequencies Barcelona might come out at, and thus make some use of this information, but it is definitely a difference from AMD's old statements.

Not really. Based on projected release dates, AMD is basically just releasing the results of simulations. They almost certainly don't have real silicon yet, and until they do, it'll be pretty hard for them to guess how fast they'll run. The design gives them a pretty fair idea of what the theoretical limit should be, but a lot less about what they can actually produce practically. Just for example, when Intel was working on Netburst, they were talking about the possibility of 10 GHz processors. Obviously, the reality fell somewhat short of that.

In addition these processors are targeted at the server market, and in the server market performance per watt often matters more than performance per system. Power consumed is generally about linear with clock speed, so in this respect a faster clock rarely does any real good. Of course, there are other contributors to power consumption, but IPC and performance per watt tend to correlate to at least some degree.

Right now, AMD undoubtedly has targets for both clock speed and power consumption, but very little real data to support any kind of announcement about either one.

Re:Nice attempt, AMD.

[Chris Burke]

Not really. Based on projected release dates, AMD is basically just releasing the results of simulations. They almost certainly don't have real silicon yet, and until they do, it'll be pretty hard for them to guess how fast they'll run.

Since the projected release date is in the next six months, I'll say it's quite certain that they have real silicon. It may not be the silicon they end up going to production with -- historically it's unlikely that rev A0 ends up being the production rev -- but they must have silicon for testing and validation by now or they'll never make the Q3 target. It can take 2-3 months to get a new mask set that changes the base layers through the fab, so if they find a single functional bug they eat up half the time until launch.

Though none of that contradicts my point -- that AMD is not talking about total system performance, they are talking about half of the performance equation. This is not what AMD has liked to talk about in the past, and is misleading to take by itself. The state of their current performance projections -- whether purely simulated or based on early silicon revs -- doesn't enter into it other than as the justification for the change.

In addition these processors are targeted at the server market, and in the server market performance per watt often matters more than performance per system. Power consumed is generally about linear with clock speed, so in this respect a faster clock rarely does any real good. Of course, there are other contributors to power consumption, but IPC and performance per watt tend to correlate to at least some degree.

Yes, performance/watt is now the metric of top importance, though if the performance is too low then you need too many machines -- low power or not -- to fit in your data center to get good enough throughput.

It's true that power scales linearly with frequency. IPC is still not a good measure of perf/watt. Higher IPC usually comes at the cost of more transistors and more transistors firing at once. Widening a data bus to be twice as wide, or doubling the width of a functional unit -- both of which Barcelona has done with the FP units -- has essentially the same effect on power as clocking the bus twice as fast. Maybe more because more transistors means more leakage, maybe less because the change is local rather than applied to the entire clock tree. Depending on how extensive the changes are, how many more transistors are added, it could come out either way.

In general, you can't say higher IPC is better for power. Now you could look at the Netburst architecture and think this isn't true, high frequency obviously sucks for power. In the limit that's true, but because to get that high frequency you need to have many pipe stages which means many more latches which have both dynamic power and static leakage power. It was the double whammy of the increasing dynamic power and the jump in static leakage current that started to become problematic at 90nm that doomed the high transistor count Prescott. But all this tells us is that 20+ stage pipelines are bad for power, and a bad idea in general. For more reasonable middle of the road designs, the tradeoff between IPC and frequency to get more perf/watt isn't as clear.

Re:Nice attempt, AMD.

[maraist]

Wouldn't it be great for every app to essentially have its own virtual world? That would indicate a much lower potential for harm inherent to the OS architecture.

Not quite there with you on this one. Take a UNIX environment for example. There are a lot of 'localhost' services which are required for a VM to be useful. It's obvious for a UI front-end, but even in the back-end you've got cron-tasks et-al. Whatever the OS or technology, these boiler-plate services consume resources, often in a wasteful way (disk-caching being a huge one).

So while I'm all about VM's, for me, they represent a way of doing more with less hardware when you physically can't put two components onto the same OS. A great use-case is testing-environments. I know people that are all about the VM-snapshot and recover on a separate machine, but there was a slash-article about the apparent 15% reduction in web-performance when running in a VM.

Maybe I'm just old and grumpy now, and my view of the VM is similar to people's view of multi-threading 15 years ago - that MT will never be as fast as sequential logic, except that you can throw massive amounts of parallel hardware at the problem. Likewise in 2 to 5 years, the same will probably be said about VM's.

Re:Nice attempt, AMD.

[trentblase]

s/fanboy/stockholder

Re:Nice attempt, AMD.

[OwnedByTwoCats]

For the last 30 years, all of semiconductors has been based on "and it'll be even cheaper in six to nine months". It has always been "buy it now, or wait six to nine months for the next cool stuff".

Re:Nice attempt, AMD.

[Chris Burke]

Intel can't copy AMD's advances. They'd have to license some tech to do so, and that's not very likely (I believe HTT for instance belongs to AMD).

Actually Hypertransport is an open standard and anyone can implement it. AMD doesn't have the clout to force proprietary standards on the market, so their only hope to have a standard adopted is to make it open and royalty free.

Which is why Intel will (probably) never implement it. They aren't interested in standards which they don't control. They already don't like the fact that AMD is cross-licensed for all x86 tech, which was part of the motivation for creating the entirely separate IA-64 ISA. When IA-64 failed and Intel was forced to implement x86-64, the only reason they used AMD's spec was because Microsoft said that they would only support one x86-64 ISA, and AMD got their first. Basically it took MS to out-monopoly Intel. So unless they are forced to use HT, they won't, and I can't see any way they could be forced. They may implement something similar -- they will have to in order to address multi-socket scalability -- but it will not be compatible.

AMD would love for Intel to copy their tech. Every time they do, it makes AMD look like the leader and Intel the follower. You could practically hear the screams of orgasmic joy from AMD when Intel announced EMT64.

Re:Nice attempt, AMD.

[level_headed_midwest]

The P6 micro-architecture of the Core 2 line is older than the PIII as it debuted in 1995 with the Pentium Pro. However, the P6 core was a solid one and was vastly reworked into the Core 2. The K10 is a derivative of the 1999 K7 Athlon, and like the Core 2, it's a very tweaked and reworked version of that.

The newest x86 architecture was killed off last year with the late 2000-era NetBurst being scrapped after it hit the thermal wall and the K8s smoked them. Newer isn't always better.

Re:Nice attempt, AMD.

[Mattsson]

"40% more IPC" isn't really any more informative than "40% more frequency". This depends.
If they say "A has 40% more performance @ 2GHz than B @ 2Ghz", comparing their relative performance is just a matter of math.
A at 2GHz will have the same performance as B at 2.8GHz.
Saying "A will have 40% higher Hz than B" and not mentioning the performance per Hz (á la the 90's) makes it impossible to make any performance estimates.

It also says a little about the performance-gains one can expect over time.
A certain chip-die will not increase it's performance/Hz ratio during it's lifespan, but it will increase in Hz.

If product A(amd) and B(intel) both gain the same percentage of clock frequency rise during a certain time-frame,
and neither introduces a new die in that time, the one with the steepest performance/Hz-curve will gain the most performance.
The same one will also gain the most from overclocking.

Re:Nice attempt, AMD.

[Anonymous Coward]

If you want to make AMD sound better, you avoid any mention of process tech.

Intel has the best manufacturing, bar none. AMD's numbers for their 45nm process
are about on par with Intel's 65nm. Not to mention that Intel is going to have
their 45 before AMD does.

Re:Nice attempt, AMD.

[Anonymous Coward]

The orgasmic screams of joy from INTEL fanboys is juiced all over Slashdot. It's disgusting.

Re:Nice attempt, AMD.

[X]

Actually, hasn't Intel had integrated L2 cache since the Pentium Pro?

Re:Nice attempt, AMD.

[Gr8Apes]

I agree with you in general. This time, however, we're looking at a huge leap in 6-9 months. Quad procs at reasonable prices is what everyone is hoping for. If we get a huge performance boost at the same time, that's just icing on the cake. In any case, I don't expect to see huge further price reductions on current lines after the last major cut.

Re:Nice attempt, AMD.

[Gr8Apes]

There was some discussion a few years ago about how the K8 core deviated from x86 internally, being more RISC than CISC allowing for greatly improved performance. I haven't seen anything similarly in-depth on the C2D's CPU internals, most articles seem to focus on the cache solution as that's apparently a new thing for x86.

From what I've read on K8L (K10, whatever it's being called this week) the internals have gotten a mixture of evolution of existing features plus a similar shared L3 cache. It's not clear if the AMD L3 cache is going to be quite as optimized as the Intel L2 cache solution, but I'd hazard a guess that it won't be.

Re:Nice attempt, AMD.

[Gr8Apes]

For some reason, I thought HTT had to be licensed. That aside, your insight as to Intel's reluctance is spot on. They've already had to eat crow thrice (I just had to use it:) this decade, with the AMD-64 implementation, the failure of Itanium, and the failure of P4. A fourth stating their entire architectural approach was incorrect might just doom them in mindshare to a complete failure.

I don't think that anyone that's truly interested in the technical aspects of CPUs thinks Intel is a leader in this decade. Since the AMD-64 and Itanium launches, that position has been firmly held by AMD which continues to extend that lead from everything I'm seeing. (For you Intel fanbois, that's even despite the C2D, which really is irrelevant in the world of servers). AMD's technical lead is pretty much cemented in the server world, and I wouldn't be surprised if Barcelona puts that market firmly out of Intel's reach. The current Opterons match Intel's best in equivalent configurations, but scale far beyond what Intel is capable of, and apparently Intel won't have an answer for at least another 12 months.

Intel has a PTP alternative to HT in the works, but it won't be ready until late 2008 (Nehalem) supposedly. By then, Barcelona will have been out for at least a year.

Re:Nice attempt, AMD.

[Gr8Apes]

I understand your statements about resources. But look at systems today. Can you honestly say that the current "normal" hardware couldn't easily be tweaked with minimal investment to easily handle a VM for your browser, email, and word processing client? That's what most common desktops run for 95+% of the populace, which also happen to account for most of the virus/worm issues. VMs could almost eliminate most of the issues associated with those vectors.

VMs do have overhead when run in software. But the new crop of CPUs will have hardware virtualization which should address most of the performance issues. I know that Sun hardware has been able to be configured like this for at least the past 8 years, and the performance hit was negligible.

Change is coming, time will tell if we accept it or not.

Re:Nice attempt, AMD.

[rshimizu12]

Now that AMD is delivering it's new chip it should be able to raise sufficient capital to develop it's next generation of chips. It's very logical that AMD's announcement conincided with the issuance of new a new bond offering... I believe we are going to see a accelerated time schedule for IBM and AMD's transition to 45nm chip technology.

Re:Nice attempt, AMD.

[default luser]

While the Core 2 has its roots in the P6 microarchitecture, it is an entiely different processor design with far too many changes to be called simply a "tweak" of the P6.

The reason you have not seen in-depth articles for the Core 2 is because most "review" sites on the web are staffed by people who don't know a lick about processor architectures, and they just regurgitate Intel's presskit material without any discussion on the subject. Here is a much more in-depth article concerning the architecture.

Summary of improvements since P6:

4-wide decoding instead of 3-wide.
Micro-op and Micro-op fusion (more instructions decoded).
Improved branch predicition, including loop detector (from Pentium M).
Support for 2 128-bit packed SSE instructions per-cycle.
Three dispatch units (P6 had two ALU / FPU ports).
Speculative reordering of loads.
And, of course, the low-latency (from Pentium M) and shared L2 cache. x86-64 support also counts for something.

Why does it take so long to make sizeable improvements on the P6 microarchitecture? The reason is complexity: every extra decode pipe, every extra issue port, every extra ALU adds exponential complexity to the design. Intel was actualy trying to get around this problem with the Pentium 4: crank up the clock speed, and you don't have to make a more complex processor!

The failure of the Pentium 4 and the complexity of massively superscalar cores is the main reason why CPU designers are moving toward multi-core as a long-term solution: superscalar architectures as complex as Core 2 (and Barcelona) are VERY difficult to design and verify.

Re:Nice attempt, AMD.

[default luser]

Actually, hasn't Intel had integrated L2 cache since the Pentium Pro?

No. The L2 cache of the Pentium Pro was on a second die. The two dies were packaged together in a somewhat similar manner to Intel's current "Quad-core" chips, except the cache attached to the brand-new back-side bus (instead of sharing the front-side bus).

The first Intel processor with integrated cache was the Mendocino-core Celeron.

Re:Nice attempt, AMD.

[bigpicture]

Even a blind person might notice that you are an impartial Intelboy. Maybe an MSboy too? Or can we say a WinTelboy? Get with the program and discuss what is really going to be relevant to the future.

Re:Nice attempt, AMD.

[Anonymous Coward]

Intel's robustness in the face of failure is actually very impressive. They never put all their eggs in one basket. That's the only sane way to run a cutting-edge company, but how many managers can really swallow the redundancy that implies? The most common strategy in tech (after the initial risk-takers retire on their stock options) is to muddle along on a technological worst-case trajectory. Innovation means risk, risk means failures, failures mean redundant backups, redundant backups mean inefficiency, inefficiency means layoffs. Obviously Intel doesn't buy into this line of reasoning, or the Core project would have been killed at birth and all those resources committed to NetBurst.

Sometimes it seems like the most basic concepts of risk and return are understood only by a few very old and elite companies like Intel and IBM. At my company, no projects fail. They just take a long time and produce disappointing results. That's because we never do anything that might fail.

Other companies have the guts to try things that might fail, but they pretend like there's no possibility of failure. They make no plans for failure, and when things fail, they always blame it on avoidable mistakes by individuals. Then they make grand plans to prevent projects from ever failing again. If they succeed, they end up like my company.

Companies should be like investors. They should identify the amount of volatility they can tolerate and invest in a mix of high-risk, high-return and low-risk, low-return projects. They should realize that eliminating high-risk projects often decreases their rate of return. Instead, they treat elimination of risk as an absolute good, no matter what the cost.

Re:Nice attempt, AMD.

[maraist]

In terms of viruses, almost all of them could be eliminated if they used a simple unprivledged user-space. Forget VM's, have each application run on as it's own unprivledged user.. Then viruses would be impossible.

The problem is that this is not useful to the user.. Windows (and Macs) exist because every application seamlessly interacts with every other application.. We're even seeing this in the gnome/KDE camp more and more..

People want to download a widget, click install, have it run as root to add to the local task-bar of all users.

This is impossible with VM's, but it's also very difficult with unprivledged users (ignore for the moment over-privledged Linux sudo access which is becoming prevalent).

Consider the windows registry (or the gnome equivalent). Every app can fuddle with every other app's data.. You have read-access to all your bookmark data, browser-history data, contact lists.. All you need to know is A) how to get the user to execute a semi-privledged process, B) the format of the application data.

Unless you change people's mindsets about application isolation, you're not going to get VM adoption.. And user-space isolation will be more resource friendly anyway.

Bring on the competition

[heinousjay]

I really hope this plays out. Not only do we benefit from better technology, but I get to read all the fanboy flamewars too!

Re:Bring on the competition

[Anonymous Coward]

Screw you! Fanboy flamewars suck!

Re:zoom!

[mdocod]

lets see the bacon

i certainly hope so

[PhrostyMcByte]

It would be a shame if after what, 4 or 5 years? of being in the lead, AMD loses focus and stops making fast CPUs.

The last thing we need is for Intel to have no real competitors. Innovation would slow and prices would hike up.

Re:i certainly hope so

[Bodrius]

It would be a shame if after what, 4 or 5 years? of being a strong competitor, AMD loses focus on what is relevant in the market and just keeps making faster, more power-hungry, CPUs.

The last thing we need is for Intel to have no real competitors. Innovation would slow and prices would hike up.

Re:i certainly hope so

[Jeff DeMaagd]

It's not as if AMD was suddenly not making fast CPUs, just that Intel's best was faster than AMD's best.

There was just a little issue of whether they make the fastest. I would expect that few people would notice the difference in speed between a top AMD and a top Intel chip in actual use.

Re:i certainly hope so

[petermgreen]

making the fastest is a big deal.

Afaict intel and amd both make significant profits from selling the cream of thier production at huge markups. It is much easier for AMD to do that if thier cream is currently better than the competitors cream (the same applies to intel but to a lesser extent because intel is the gorilla with lots of contracts and marketing).

Re:i certainly hope so

[Jeff DeMaagd]

I think that makes sense, but still, AMD survived a very long time by being "second best", I really don't see it to be a major problem, except for the short-term damage caused by the ATI merger. Only time will tell if they made a good choice for the long term. A lot of people seemed to like AMD for having better bang for the buck.

SPECint_rate_2006 vs SPECint_rate2006?

[FooAtWFU]

Wow! The underscore makes all the difference!

Yes!!!

[RingDev]

I just bought an Intel Core 2 Duo, and I love it. But I was beginning to worry that AMD's rocky quarter, lack-luster product line up, and soon to be cut backs, might lead to a less competitive playing field. But I'm excited to hear that AMD is still in this fight and will be upping the ante for my next PC purchase.

-Rick

Re:Yes!!!

[geekoid]

Historically, company come out with something 'unexpected and amazing' after a really disasterious quarter.
I would be prudent.

Using "up to" in benchmarks and comparisons...

[jimstapleton]

has always bothered me.

"Up to" is sugar-coated for "You can't expect any better than this" with a implicit translation of "It can get a whole lot worse".

Ex: If CPU X get "up to" 100% more performance than CPU Y, but in all tests but one, actually has 1% of the performance, I'd rather have CPU Y.

"Up to" means nothing to me, except as an advertisement for the competator; whichever has the least unpleasant average and worst case performance is the one I'm interested in.

Re:Using "up to" in benchmarks and comparisons...

[grasshoppa]

"Up to" means nothing to me, except as an advertisement for the competator; whichever has the least unpleasant average and worst case performance is the one I'm interested in.

And those numbers would be indicative of anything either. The problem with CPU benchmarks is that they have no real world application; Everyone has different needs. However, the marketing types for both the suppliers and consumers need numbers to push in front of each other, so they make up these things which those of us in the field understand have no real world meaning.

It's a vicious circle, non?

Re:Using "up to" in benchmarks and comparisons...

[Smidge204]

Actually, benchmarks DO have some real-world meaning - but only for comparison.

If your specific needs happen to be similar to the things benchmarks stress, then you can expect the results to be relevant. If your needs differ wildly from benchmark methods, then you can expect the results to be irrelevant - but most likely they will be equally irrelevant.

Benchmark performance is, at the very least, a better indication of relative performance than clock speed of cache size.

Fact is, few end users actually NEED the kind of power modern processors provide. How often does your typical web surfing, middle class consumer see a CPU usage above even 10%? Most of that power goes into running heavier and heavier GUIs and OSs instead of actual work anyway. Anyone who would actually use a machine for all it's worth will probably know enouh about computer systems to know if a benchmark rating is useful for them or not.
=Smidge=

Re:Using "up to" in benchmarks and comparisons...

[donglekey]

Not true. I have a top of the line computer from 3.5 years ago, and it cannot play high def trailers. Combine this with lots of flash video, video conferencing, etc. and you have various cpu needs that need to be met. It is hard for me to believe too, but flash and quicktime are pushing the needs of people's hardware.

Re:Using "up to" in benchmarks and comparisons...

[seaturnip]

How often does your typical web surfing, middle class consumer see a CPU usage above even 10%?

When I load a large slashdot page in Firefox, my CPU usage shoots to 100% for several seconds.

Re:Using "up to" in benchmarks and comparisons...

[ponos]

Not true. I have a top of the line computer from 3.5 years ago, and it cannot play high def trailers. Combine this with lots of flash video, video conferencing, etc. and you have various cpu needs that need to be met. It is hard for me to believe too, but flash and quicktime are pushing the needs of people's hardware.

This is an indication that the code sucks. You should be able to play almost anything with a processor rated as 3000+ or better. I am sure that you can find some unrealistic 1080p/MPEG4 content that can stress some modern processors, but as a general rule, I haven't encountered ANY video that stresses my athlon 3200+ with mplayer (linux). Maybe you need a better video card?

P.

Re:Using "up to" in benchmarks and comparisons...

[NatteringNabob]

The programs that make up SPEC *are* real world programs that have been selected by the SPEC committee as representative of typical workloads. The mix of programs that make up SPEC change over time to reflect the fact that CPU's become more capable. In particular, many older benchmarks in the SPEC suite have been dropped because they now fit entirely in L2 cache.

All that said, this is the hardware equivalent of Fredrick Brooks maxim that 'the Paper tiger is always better than the actual one unless reality is required'. What matters is what AMD is actually delivering at what price at what volume. A chip that you can't buy has an effective real world SPEC_int and SPEC_fp of 0.

Re:Using "up to" in benchmarks and comparisons...

[Anonymous Coward]

I have a sort of top of the line notebook from 3+ years ago and it *CAN* play high def trailers.

Buy a better video card and you'll be all set :)

Re:Using "up to" in benchmarks and comparisons...

[Gr8Apes]

You must be using FF 2.x.

I must say that my experiences with 2.x are making me fall back to 1.5.x. There's something rotten with the 2.x release, and I'm not sure what it is, but the fact that I can outtype the system in a text area while responding to /. means that either I'm the world's fastest typist or that something's seriously wrong with this 2.x release I'm using right now. (I should note that spell checking has been turned off, but that has not eliminated the problem. My home system with no plugins (that's right, I went that far to browse /. with no adblock, flashblock, noscript etc) also has hesitations that 1.5x does not have.

Re:Using "up to" in benchmarks and comparisons...

[donglekey]

Athalon 2700, gf6800, 720p Quicktimes run at about 10 fps, don't know why. Pretty fresh install of WinXP and I don't use IE, but spyware never seems to outside the realm of suspects.

Re:Using "up to" in benchmarks and comparisons...

[general_re]

Athalon 2700, gf6800, 720p Quicktimes run at about 10 fps, don't know why. Pretty fresh install of WinXP and I don't use IE, but spyware never seems to outside the realm of suspects. It's QT - I've noticed the same problem on a 2.8Ghz P4. However, I can play 720p WMV encoded HD content just fine on the same machine. Try it yourself:

http://www.microsoft.com/windows/windowsmedia/musi candvideo/hdvideo/contentshowcase.aspx

Re:Using "up to" in benchmarks and comparisons...

[ponos]

Athlon 2700, gf6800, 720p Quicktimes run at about 10 fps, don't know why. Pretty fresh install of WinXP and I don't use IE, but spyware never seems to outside the realm of suspects.

If possible, try saving the videos and running them from a local folder with a decent player, like VLC. The in-browser players usually suck (performance-wise). You machine is definitely capable of running 720p content, even with relatively modern codecs.

P.

Re:Using "up to" in benchmarks and comparisons...

[donglekey]

Thank you to both posts. I will try it, Viva VLC.

"at the same frequency" is pointless

[Wesley Felter]

When the fastest Barcelona is ~2.5GHz and Clovertown is 3.0GHz, comparisons at the same frequency are pointless. What matters in reality is performance at the same price or performance at the same power or highest available performance at any price.

Re:"at the same frequency" is pointless

[Courageous]

Agree, caveat that performance-per-watt needs to be added in, at least in part because watts are money.

C//

Re:"at the same frequency" is pointless

[Gr8Apes]

So AMD will win outright then?

Seriously, Intel's got nothing on AMD in performance per watt.

Re:"at the same frequency" is pointless

[Anonymous Coward]

huh? I'm not sure if you've been keeping up with recent hardware developments. Ever since the Core2Duo, Intel's taken the performance-per--watt crown from AMD.

See: http://www.anandtech.com/cpuchipsets/showdoc.aspx? i=2933&p=9

Re:"at the same frequency" is pointless

[Gr8Apes]

that article happens to state that
1) only the 3800 is an EE chip.
2) they're running on one of the most power hungry motherboard chipsets made for AMD: the nVidia 590 SLI
3) only the X2 5000+ is a 65 nm CPU

So basically, let's stack the deck as much as possible against AMD in this test without showing a best case scenario, while postulating that they're showing a "worst-case scenario" with a "bad E6300 sample".

I like Anandtech usually, but this article could almost have been sponsored by Intel and is far from Anand's usual quality and attention to detail. It almost reads like a Tom's Hardware piece. Lines like the below from the Conclusion just cement that feeling. So equal prices and equal performance? (Remember, AMD's tech is 3 years old...) Oh, and AMD is faster in "outlier cases". Intel should have totally rocked these tests. That they don't indicate something else. Lastly, let's note that Intel has already stated a large price cut for the fall, around when AMD is set to ship. I wonder if they got hold of a pre-prod Barcelona or two and had to change their undershorts?

With the latest round of price cuts AMD is far more competitive than at any other point since the release of Intel's Core 2 processors. Unfortunately for AMD, this means that at best, it can offer performance close to that of Intel's Core 2 processors at similar prices.

Overall, the performance advantage still goes to Intel's Core 2 lineup but there are a few situations where the performance between the two families is close enough to be considered a tie. There are also the outlier cases where the Athlon 64 X2 actually ends up faster than the Core 2, but we suspect that they are more isolated incidents than indications of the norm.

Re:"at the same frequency" is pointless

[Courageous]

I agree with all that. IIRC, advantage is to AMD for idle power consumption also; OTOH, with virtualization on the upswing, CPU's are less and less idle these days.

Speaking of virtualization, AMD seems to be doing the right thing there... nested page tables and so forth, along with the L2/L3 cache combo which they tested out as being better for virtual environments.

AMD has an advantage over Intel at any given process size, on the grounds that the IBM-AMD SOI process is basically better at any given process size...

I'm definitely looking forward to Barcelona and near-family releases from AMD. Should be interesting.

Intel pulling their head out of their ass was interesting, too. If they hadn't of, they'd be facing a generalized collapse of their market right now. AMD was utterly killing them.

C//

Re:"at the same frequency" is pointless

[Gr8Apes]

Intel pulling their head out of their ass was interesting, too. If they hadn't of, they'd be facing a generalized collapse of their market right now. AMD was utterly killing them. Yes, that was interesting :) I'd make the argument that Intel already has a collapsed market: servers. Take a look what any halfway competent IT manager is buying for x86 server hardware. It's almost all Opterons. The server market collapse should continue with the consumer market shortly, thanks to several monopoly lawsuits coming to a close and just much better hardware from AMD. By collapse I merely mean the collapse of Intel's monopoly, a welcome thing. I don't for a second believe Intel will dissappear, although that's almost the case in the server market. I can only hope that Apple will add AMD processors to their lineup in the near future, even if it's only in their servers/Mac Pros.

Re:"at the same frequency" is pointless

[Courageous]

"It's almost all Opterons." Source?

C//

Re:"at the same frequency" is pointless

[Gr8Apes]

From Anand, AMD, News.com and X-bit labs the numbers tell of an ever increasing marketshare for AMD. Note that this is for all servers, not just against Intel, meaning that Sun, IBM, and other servers are included. Then there's the 4 socket server category, where AMD has 48% of total marketshare in Q2 2006, and is the category I'd be more interested in as that is a true high-end server market. Then there's the Top 500 Nov 2006 List which lists 4 Opteron systems and a single Intel Itanium system in the top 10, along with 5 PowerPC systems. I should also note that only 31 Woodcrest systems are on the list vs 76 dual core Opterons.

Basically, AMD's server market share has been growing in leaps and bounds over the past 2 years, and has broken Intel's x86 monopoly in the space, especially once you exceed 2 processors. With Dell now finally offering AMD CPUs, I expect that number to grow.

Re:"at the same frequency" is pointless

[Courageous]

I see. You were speaking figuratively when you said "almost all".

Their peak category is 48% of 4 socket boxen, at 48%, which would be "almost half,"
not "almost all," and I agree, because of HT they are going to own >= 4 socket space.

Woodcrest systems are selling out left and right, though, and the 4 *CORE* space
is a bigger space then the 8 core (4 socket) space, with more total sales. Every blade
except Sun's is 4 cores (and the Sun solution is specialized towards ridiculously
high IO and is therefore very expensive).

C//

Re:"at the same frequency" is pointless

[Gr8Apes]

In the 4 socket space, they are 48% of the entire 4 socket market, not just x86. I couldn't find a statistic on x86 alone.

As for 1-8 core market (pre-quad core) AMD had 26+% of the x86 market in either Q3 or Q4 of 2006. That's significant enough to break Intel's hold on the market. I couldn't find a statement about the 2 socket market as a whole or x86 alone, so I cannot break down the numbers any further than that. It was also without Dell, who sells about 1 out of every 5 PCs, although I don't know what their server numbers are.

However, if you realize that prior to 2004 AMD owned virtually 0% of that market....

I also haven't heard of a Woodcrest shortage. Everything I've seen in that space says that except for equivalent single socket solutions, Opterons are the better buy, especially considering the upgrade path.

Re:"at the same frequency" is pointless

[Courageous]

I'm a little confused why you think that ("Opterons are the better buy"). I personally think that the 51XX series is really rompin'. Granted, I haven't twiddled my trade for some months... have the price changes really changed the price-performance ratio? I do know that at the time the 5160 first hit the market, it was king for $$$/unit-of-compute, as well as $$$/watt continuous power. Idle was still advantage AMD, mind.

But go ahead and elaborate. I conduct trade studies for a major corporation, have input into a large data center. Am always happy to hear various opinions...

C//

Re:"at the same frequency" is pointless

[Gr8Apes]

Opterons are upgradable with a chip replacement. Would you do this in a data center for x86? I have in the distant past but within the past 8 years, replacement made more sense, since most were 1 or 2P systems. For 4P or more it used to be specialized motherboards and chipsets that couldn't be replaced (in the x86 world). With Opterons, you can upgrade CPUs.

Having said all that, if you're never going to upgrade data center systems I'd have to evaluate Intel vs Opteron for 1 or 2P systems as I haven't done so in 6 months since the Woodcrest set came out with one exception. I noted that in heavily multi-threaded applications running across 4 cores with high memory bandwidth requirements that Woodcrest systems falter in performance (limits of the FSB). Also, the FBDIMMs required on 2P 51xx systems have much higher latency. It's highly depedendent on your application whether this will make or break the decision for you. If your system benefits from NUMA, then Opterons are the only way to go. It's much more complicated for me than a simple $$$/unit-compute or $$$/watt issue. It's a $$$/performance issue, where performance is full system based.

I'll refrain from talking about Barcelona at this point, since without benchmarked processors it's largely speculative as to the various things it will bring to the table that could change the equations.

Re:"at the same frequency" is pointless

[Courageous]

For our most recent trade, what we looked at was blade systems holistically. We do image processing... we need both lots of compute power and reams of I/O. Peak IO was held by Sun, for their 8000s, which if you look, you'll find to be ridiculously equipped for IO. However, we were somewhat cost sensitive, and went more for a CPU/$$$, IO/$$$, with some additional sensitivities towards footprint and power and the like. Dell's 1955 blade was the winner here.

Sun's 8 core blades are really impressive. However, we're moving to a scale-out environment where larger SMP boxen aren't so important. Stateless SOA, blah, blah, you get me I'm sure.

C//

Re:"at the same frequency" is pointless

[Gr8Apes]

very interesting application, and I'm guessing you have (a) controller systems that segment and hand out work to individual systems/processors for the Intel set to work more efficiently than AMD (given that 2P 51xx systems start to have memory bandwidth issues if multiple threads access shared memory across both processors - the same problem that exist with the quads, btw)

The most ridiculously equipped IO system I've seen are SGI systems, but that was years ago and haven't revisted for that level of bandwidth since, not being in the same market anymore.

Interesting discussion. I'm likewise about to setup a configuration for SOA/HA, but for data services which are slightly less, shall we say, IO intensive. ;)

Correction

[matt me]

Barcelona will have a 50% advantage over Clovertown in floating point applications and 20% in integer performance. I think the figures for relative performance should be 1.500000000000000 and 1 respectively.

Re:Correction

[Minwee]

I think the figures for relative performance should be 1.500000000000000 and 1 respectively.

Of course Intel claims that they are 1.499999999326112 and 0.999999994351582.

amd quad-cores are true quad-cores with a better..

[Joe The Dragon]

Chipset to cpu and cpu to cpu link with intel you have to use the chip set for one cpu to talk to another one.
Also If amd where to copy intel and put 2 dies on the same cpu they will have a better link for them that will not eat up chipset to cpu bandwith.

Why casual users can't be bothered

[ushering05401]

Over the past week we have heard about Intel's dominance and flashy new products, AMD's disastrous quarter, and now AMD's supposedly dominant new offering.

I read tech news daily and am getting sick of the media wars... It is no wonder casual users get fatigued trying to keep up. Casual opinions depend on which day (or week or month) a person chooses to research product offerings. It is no wonder I am always hitting a brick wall when trying to get my users to educate themselves so they can get more out of their tech. They don't know what to make of all the posturing.

This is not a function of the tech world developing *that* quickly. It is a result of the major players trying to out-strategize each other. I don't want to see anymore benchmarks (or hear about anymore promised software) until I am standing in front of a demo machine that is running the tech.

Guess I woke up on the wrong side of the bed.

Regards.

Re:Why casual users can't be bothered

[Gr8Apes]

Well, blame Intel. ;)

Seriously though, Intel's got the performance lead for now, but AMD's got the better tech and their release schedule "lags" Intel's a little. So Intel got the "jump" on AMD release cycle wise, and now you've got the situation where Intel has a brand spanking new product out that beats AMD's old offering by about 10-20%, at best at stock speeds.

I personally am waiting for AMD's release and benchmarks before making a final decision, but the fact that I'm doing so already says which way I'm leaning. I should also mention I recently bought both a C2D and an X2 system, so I'm not exactly a fanboi in either direction. The C2D is in a laptop and rocks. The X2 is in a desktop and OC'd and for the money an Intel C2D system couldn't touch it. I had 2GB of DDR RAM sitting on my desk, so that's essentially free and a $40 high end motherboard didn't hurt things either.

Re:Why casual users can't be bothered

[networkBoy]

FWIW, it seems to be AMD doing all the posturing.
Intel seems to have taken a "no response" approach to media claims, instead producing product and letting guys like toms hardware do their thing. This isn't to say they don't advertise, but they don't take out full page NYT (or was it washington post?) ads chest pounding like AMD does.

-nB

Re:Why casual users can't be bothered

[networkBoy]

if by lags you mean 3 year spread...
-nB

Re:Why casual users can't be bothered

[Anonymous Coward]

FWIW, it seems to be the Democrats doing all the posturing.
Republicans seems to have taken a "no response" approach to media claims, instead running the countryt and letting guys like bill O'Reilly do their thing. This isn't to say they don't advertise, but they don't take out full page NYT (or was it washington post?) ads chest pounding like the Democrats do.

Re:Why casual users can't be bothered

[Chandon Seldon]

Wait... what?

That's why we're constantly hearing about the performance advantages of Penryn. If anything, AMD has been a bit quieter than Intel. Compare the references on the Wikipedia pages for Barcelona and Penryn if you want evidence.

Re:Why casual users can't be bothered

[logicnazi]

Well, as much as I think Intel usually gets a bad rap on slashdot and similar places, in fairness I out to point out that this is really easy to do when you have the performance crown. Taking out ads bragging about their superior performance would mostly just give people a reason to doubt intel (if they are taking out ads does it mean it's in doubt?) while AMD had better take out these ads as at the moment no one else is going to do it for them.

If the situation ever reverses and AMD's strategy to keep up sales while on the losing end works don't be surprised for Intel to do exactly the same thing.

Re:Why casual users can't be bothered

[codemachine]

I think this has more to do with the fact that "Intel Inside" and such have been ingrained in people from Intel's past advertising. The general public is much more likely to have heard of Intel than AMD, which means AMD has a much greater need to get their name out there than Intel.

Re:Why casual users can't be bothered

[Bo'Bob'O]

Well, fact is, when it comes to CPUs, in the end, whether Athalon, Sempron, X2, Core 2, Celeron.. all these things over the last decade, if you spent X on one product (not including server chips) the difference in performance was generally measured in a few percent.

On the other hand, buying video cards has gotten BAD. The numbers, ratings, even the price points have become seemingly almost random. The mark up is enormous at just about any retail chain. Higher price cards can have lower performance sometimes even in the same series. Most review sites can't even keep up, and even if they do, reviews are often spread out over multiple articles, making A/B comparisons extreamly difficult. It's even worse for laptop and integrated graphics. Unless you keep up constantly, it's difficult to make even casual recommendations.

How much for HyperTransport?

[Anonymous Coward]

We already know AMD at 4-cores beats Intel. Its not due to processor design, but the mere fact that Intel can't feed its processors fast enough. AMD's HyperTransport provides the bandwidth, while their integrated memory controller helps hide the latency. Intel instead has traditionally favored larger and smarter caches to resolve this, but this doesn't scale as well. Their next generation will put them on even or better terms than AMD.

So, the real question is how the cores compare head-to-head? We need to know where this supposed gain is coming from, which will tell us how far behind (or ahead) Intel is.

Re:How much for HyperTransport?

[dfghjk]

"We already know AMD at 4-cores beats Intel."

Who is "we" that already "know"s that? What 4-core AMD processor beats its Intel competition?

Re:How much for HyperTransport?

[Anonymous Coward]

Anyone whose looked at the benchmarks. :)

Yes, a 4-core procesor beats Intel's. This isn't due to AMD having a "true" 4-core (utter bs, btw), but due to bandwidth. This will likely change once Intel releases its next generation infrustructure (CSI and on-chip memory controller). Until then, AMD can continue to raise 4-core performance for PR, while their singe and dual core varients continue to be outpaced by Intel. Thus, the question of whether this is a true architectural improvement (only minor in K8L, from what I've heard) or largely based by exploiting HT (which means that by the time the chips are released, they'll be slower then Intel's).

Re:How much for HyperTransport?

[logicnazi]

Well, I highly doubt that AMD's integrated memory controller's ability to hide the latency can scale any faster than Intel's cache tech. After all they basically control latency in the same manner, smartly guessing what memory the chip is going to need and making sure they keep the cache filled with that data. Of course the on board memory controller means that the latency starts at a lower place in the first place. Though it could be that the extra latency of going across the northbridge gets worse as clock speeds increase but I kinda expect it will remain proportional to bus speed.

The real issue in the next generation is simply the raw bandwidth. If AMD keeps moving in the direction it has it's next systems will be NUMA architectures like the Quad FX. This means that each chip (maybe next gen it will be core) has a constant memory bandwidth while Intel has to divide one FSB by the total number of chips in the system. Also AMD no doubt will have copied some of Intel's smart cache techniques.

Still if Intel can just keep parity (and they have the advantage of just now moving to 45nm) for another year or so (and not fuck up CSI again) that's pretty much good enough. At that point Intel no loner suffers a fundamental architectural deficit compared to AMD and their superior process technology gives them a serious advantage.

The three two questions are these:

1) How much market share can AMD gain capitalizing on Intel's memory starvation in the next year or two? Mostly this is going to depend on how well AMD can copy Intel's fancy cache and memory conflict checking hardware. If they can do a fair job of this then their underlying architectural advantages should let them overcome Intel's process advantages and make a fair bit of headway.

Unfortunately for AMD they could end up not being able to take full advantage of their superior (large scale) architecture if OS vendors don't provide better NUMA support. Eventually this is the way both AMD and Intel are going but better support now would be a huge boon for AMD.

2) Can AMD pull another rabbit out of it's hat that will give it a fundamental *architectural* advantage before Intel comes out with an acceptable version of CSI? Simply upping the GHz on HyperTransport isn't enough. AMD and Intel face the same fundamental obstacles to making high speed serial links and this area just isn't complex enough to let them totally out do Intel on transport speed. This is what AMD needs to overcome their process disadvantages and remain a serious contender for the performance and performance/watt crown (unless Intel just gives up the former to pursue the later). Given AMD's history they very well might pull this off but it's going to be tough.

Of course there is always the chance of a total shakeup in this industry if either Intel or IBM manages to patent some amazing process trick that isn't easily copied.

Re:How much for HyperTransport?

[init100]

Unfortunately for AMD they could end up not being able to take full advantage of their superior (large scale) architecture if OS vendors don't provide better NUMA support.

If there is any problem with the NUMA code in Linux, AMD is welcome to fix those problems themselves, instead of waiting for someone else to do their job for them.

vaporware

[Anonymous Coward]

Ship some engineering samples to Tom or Anand, otherwise this is just marketing BS.

Re:vaporware

[rootofevil]

reviews posted by tom are also marketing BS.

Re:vaporware

[Gr8Apes]

pre-production runs have been shipped. I'm sure we'll be starting to see some numbers sometime soon, but I feel safe in saying that AMD's increased optimism about Barcelona's performance is based on these first chips performance.

Re:vaporware

[fitten]

I agree... proof or STFU AMD. I'd be more than happy with a faster processor, I don't care who makes it. But I'm not going to buy into the hype until I see some benchmarks done with it, even the kind Intel did with Conroe and Penryn where they provided machines and babysat the reviewers.

On the other hand, with this kind of hype they're putting out, they'd better deliver or they're sunk.

Awesome!

[xerent_sweden]

I can't wait to run Microsoft Word on these babies!

Re:Awesome!

[Heembo]

I think you are the first person in the world to ever use that sentence.

((x*12)+5)/10

[Anonymous Coward]

((x*12)+5)/10

AMD needs to rebrand itself too

[Prien715]

I think one of the major reasons why AMD did so poorly last quarter was its silly marketing campaign. Towering signs on billboards and large airport ads tout AMD as "smarter choice", since it uses less power.

Marketing a chip as using less power is the same as having Toyota make an exclusive advertising campaign toward wheel-chair bound people: the group you're targeting has few people in it and they're going to research any product they buy. The server market is important, but when I buy my shiny new server, power consumption isn't my first consideration, nor is that the only thing AMD offers.

With this announcement, I'm hoping AMD starts a new slogan touting, say speed. That's what I buy a processor for primarily. AMD's always been fast for the cost and it's high time they market themselves as being faster and better rather than being "as good as" Intel. My new pick for a marketing slogan? "Upgrade to AMD" AMD should position its chips to be slightly more expensive at every pricing tier, but in doing so, blow them away in performance. (In the present economy, businesses have money and will gladly spend more money on products they feel are superior. Ford spends more money on marketing than BMW (but which would you rather own?). AMD should be trying to make Intel look like Ford, rather than being the "Ford alternative".)

AMD is marketing to a minor concern of a niche audience, while they ought to be using their superior performance (at a given price point) to sell hardware. Would you rather be a "power saver" or "upgrade to AMD".

Re:AMD needs to rebrand itself too

[Anonymous Coward]

What are you blabbing about? We all want to pay less for energy bills, so less power is a very good selling point.
It does make a difference for business. Now if you are merely speaking of home use, but even there, I want power for less cost too.

Re:AMD needs to rebrand itself too

[bockelboy]

The server market is important, but when I buy my shiny new server, power consumption isn't my first consideration, nor is that the only thing AMD offers.

That's nice - but when we look at purchasing $250k - $500k of servers, power consumption as an important factor.

Back in the days when dual-cores were just beginning, this indeed was HUGE. Do you want 30% more Irwindales which would require 100 tons of cooling, or the AMD dual-cores which require 30 tons of cooling? The same is going to happen at the dual-core/quad-core boundary.

As CPUs are cheaper and cheaper and A/C systems remain a constant cost, the people who spend large amounts of money are going to look more and more at power costs. They're probably aiming at business customers who don't buy *a* server, but buy a *hundred* servers.

Re:AMD needs to rebrand itself too

[bfields]

Now if you are merely speaking of home use, but even there, I want power for less cost too.

And I assume there's a strong correlation between power and noise. Most people's homes are quiet enough for a loud machine to really call attention to itself.

Re:AMD needs to rebrand itself too

[Anonymous Coward]

I think one of the major reasons why AMD did so poorly last quarter was its silly marketing campaign. The only thing silly here are your twisted and completely wrong car analogies. Hell, everything you just said is wrong.

Re:AMD needs to rebrand itself too

[Spoke]

Exactly. When you have a datacenter (or even a room) full of servers, the amount of heat you have to dissipate is very important.

Electricity and cooling costs are huge. And unlike a server which you buy once and may use for 3-5 years, you pay for electricity and cooling all the time. The electricy/cooling costs over the lifetime of a product can often cost more than the server itself, so anyone not looking at the power consumption of their systems as a high priority item (desktop, server, anything) is doing themselves a disservice.

Re:AMD needs to rebrand itself too

[Penguin's Advocate]

Personally, I'd rather be a "power saver"

Re:AMD needs to rebrand itself too

[Chris Burke]

Sorry, but performance/watt has become one of the major metrics by which CPUs are chosen, especially in business and double especially in business server rooms. Performance/watt is the new performance/$, because wattage tells you how much the hardware is going to cost on a continual basis for both the electricity to run it and for the necessary air conditioning. These costs dominate over the initial cost of the hardware, and are thus more important. Plus, if you have limited space you have a limited heat capacity, and higher perf/watt means you can get more perf in your server room.

Re:AMD needs to rebrand itself too

[lordSaurontheGreat]

I think one of the major reasons why AMD did so poorly last quarter was its silly marketing campaign. Towering signs on billboards and large airport ads tout AMD as "smarter choice", since it uses less power.

Marketing a chip as using less power is the same as having Toyota make an exclusive advertising campaign toward wheel-chair bound people: the group you're targeting has few people in it and they're going to research any product they buy. The server market is important, but when I buy my shiny new server, power consumption isn't my first consideration, nor is that the only thing AMD offers.
Exactly. The people who actually need lower power consumption in their chips know where to go find the data sheets to do it. The ads that AMD marketed were horrible.

Re:AMD needs to rebrand itself too

[Prien715]

My point was that the metrics are important, but the people buying the servers are going to do research regardless of the marketing material. If AMD marketed itself as "better" you'd consider it just as if they marketed themselves as "power-saving". While AMD may gain some traction in the server farm market from the ads, the waste the opportunity to gain traction in any other market by targeting a niche, and in doing so make people who want to buy AMD chips for other reasons think twice -- "You can have an Intel chip or another chip that saves power." "I'll take the regular intel chip"

AMD should market to their entire market.

Re:AMD needs to rebrand itself too

[darkwhite]

Would you rather be a "power saver" or "upgrade to AMD". Be a power saver.

When Intel came out with Centrino, I bought one almost at launch. When AMD came out with Winchester, the Athlon 64 that made the gigantic leap in price/performance/watt, I bought two. When nVidia started making lower-clocked GPUs that didn't need a fan and wiped the floor with ATI in price/performance/watt, I bought three over 3 years (6200, 7600, and now, 8500) (that was the main reason - the other one being ATI's shitty drivers for Linux). Now, I'm looking at a new ultraportable, and AMD doesn't even offer anything to match Intel's ULV parts.

Do you know why my opinion counts? Because I just helped procure a $180K HPC cluster purchase with 300 cores total. You bet your ass I paid attention to performance/watt.

(Ironically, at our specs, AMD has about 20% less power draw per U... but it also has 50% lower core density.)

Re:AMD needs to rebrand itself too

[Chris Burke]

My point was that the metrics are important, but the people buying the servers are going to do research regardless of the marketing material.

Tee hee hee! Of course, that's why there's all those ads in computer trade rags and the wall street journal, because the people buying servers always do their own research and never believe marketing materials.

AMD should market to their entire market.

Based on what? They can't exactly claim top performance right now. They can't claim longest battery life for laptops. They might be able to go after perf/dollar, but being mostly fab limited it helps them more to get higher margins, and Intel is being aggressive on pricing which doesn't leave them much room anyway.

AMD is playing to their strengths, and right now their biggest strength is perf/watt and the server market. If I'm not mistaken, the big full page add you are referencing has a picture of an Opteron -- not Athlon -- right in the middle, and refers to a server benchmark at the bottom.

Re:AMD needs to rebrand itself too

[Jesus_666]

Wait a second... So you want AMD to miraculously make their CPUs faster at any cost (they recently tried to, with less than impressive results) and additionally raise the prices by an amount not prroportionate to the speed gains? AMD took and held its chunk of the home user market by being cheaper and/or better then Intel. Once they start having worse pricing (and also horrible efficiency) it's time to consider jumping ship.

Also, forgoing power (and thus heat) efficiency isn't going to make them any friends in datacenters, as has already been pointed out. The home user market is not much better. People don't want space heaters, they want chips that barely reach 70C at full load. Not only do they pose less of a cooling problem, they also draw lesss power, resulting in a lower electricity bill. Those are very good reasons against buying a "performance at any cost" design.

I remember the Thunderbird. You could fry eggs and bacon on it without the chip dying. That was cool. You could also melt the tubing of your water cooling system or completely scre up the processor socket with it. That was less cool.


Effectively, your campaign boils down to having AMD: It's more expensive as the next AMD marketing campaign. Sorry, but efficiency is as important in a computer as in a car - and to be honest, in Germany the stereotype dictates that BMWs are only driven by wealthy assholes (or assholes trying to look wealthy) and adolescents trying to look cooler than they are (their traditional vehicle being the 3 Series).

By the way, "performance at any cost" is awfully reminiscent of 3Dfx...

Re:AMD needs to rebrand itself too

[afidel]

Not really, a DL360G5 will cost ~$1,800 over 3 years including power, the fractional percentage of the generator and the fractional part of the AC system purchase. That for a server that costs $5,800.

Heh.

[Skadet]

Indeed. I work in a law office as a graphic designer/web designer/video editor. That's what I do all day (when I'm not reading slashdot).

2 of our attorneys just got quad-core Mac Pros with Studio displays. For writing documents on. Maybe the occasional slide show. I'm stuck on this 3-year-old Dell with dual CRT monitors. Old ones.

Sorry, just had to bitch a little. Your comment is more real-world than you may have realized.

Re:Heh.

[xerent_sweden]

Yes, sadly, I have to agree to that. Then again, who knows what the next version of Word will require to display a blank paper?

Re:Heh.

[Anonymous Coward]

Keep bitching.

That and not trying to forge forward in your life will keep you right where you are.

I fucking hate people who whine about shit ass jobs they have. Learn how to sell and start earning what your worth.

Re:Heh.

[stevesliva]

2 of our attorneys just got quad-core Mac Pros with Studio displays. For writing documents on.

They probably also got Ferraris. To commute to work in.

Re:Heh.

[Skadet]

[...] Learn how to sell and start earning what your [sic] worth.

I hate selling. I'm damn good at it, but I hate it. To quote Zoidberg:

It's all so complicated with the flowers and the romance and the lies upon lies!

Selling's about kissing ass and pushing off whatever it is you're selling on whoever has a wallet, no matter what their needs. Ok, at the retail level maybe not so much. But any sales job that pays close to 6 figures, yep.

Yeah, I'll pass. And I'll be doing what I want when I leave work at 4:30pm sharp, while you're closing a deal (still) at 9pm.

Re:Heh.

[yoshi_mon]

I think the bitch was more that the budget was being used in such a way that it was not doing any good for anyone. Partner X & Y don't need five grand machines to write up their depositions on. Instead a better use of the budget would have been to buy them some nice three grand machines and spend the other 4 grand on an upgrade for someone who's work relies more on the amount of computing power they have than those attorneys.

But of course you just wanted to troll. I wonder how much you have invested in your rig such that you can troll at maximum speed?

Re:Heh.

[greg_barton]

Dude, convince them to get you a cheap mac mini and set up Xgrid clustering on those quad core mac pros. You'll be leaching off their processors in no time. :)

Re:Heh.

[Anonymous Coward]

He's a whiny bitch. He is earning what he's worth.

Re:Heh.

[Achromatic1978]

Last time I worked at a law firm, partners didn't write depositions. They signed depositions they had associates write. On the rare instance they had to "write" something themself, they did no such thing. They dictated and handed a tape to their secretary.

Re:Heh.

[kjart]

Selling's about kissing ass and pushing off whatever it is you're selling on whoever has a wallet, no matter what their needs. Ok, at the retail level maybe not so much. But any sales job that pays close to 6 figures, yep.

I'm pretty sure that he meant to learn how to sell yourself so that you can get the job that you deserve. Everyone needs to know how to do that to be successful.

copmutronium

[Anonymous Coward]

Which one is faster is pointless. They are all wastes of space as long as they confine their chip production facilities to earth's gravity and atmosphere.

DISMANTLE MARS ALREADY

Fly me...

[camperslo]

I'll do you 50% faster and 20% harder than your date last week, and promise not to cost you more.

But marry me soon baby, I need the money

SSE4? Please, don't get distracted over little things like whether or not I can cook!

Re:Fly me...

[ScriptedReplay]

SSE4? Please, don't get distracted over little things like whether or not I can cook!

SSE4? I'm not buyin' either AMD or Intel until they're at least at SSE256. What's that? It'll take a while? That's OK, I don't have the monies to get them now anyway.
<sarcasm/>

For my type of workloads, straight SSE2 is still just fine. I'll take an improvement on that now instead of, say, waiting for the x86 world to match AltiVec instruction-per-instruction. But i would go for a wider ISA - give me 4x64bit registers with the ability to do 2x128 long double calculations in parallel and I'll be all over it. Heck, even long double on the current 128bit SSE registers would be a treat. SSE4? Fits some folks' needs, but it's mostly meh! for me.

To each his/her own, I guess.

what marketroids didn't tell us ..

[appelsiini]

..it will also use 100% more wattage and produce 200% more heat than any Intel product. It will be released soon(tm), which means in few months, nobody just doesn't know which year. Good news for everybody who liked AMD Athlon thunderbird, it'll also warm up your house during cold winter months. Bad news for your electricity bill. Bind it with AMD-ATI graphics product and you'll get markets 2nd fastest rig, 'n small nuclear power plant to feed it as bonus. :-) :-)

Re:what marketroids didn't tell us ..

[Penguin's Advocate]

O the irony. Go back to the days of the P4 and you could make the same post replacing "AMD" with "Intel". Intel has only been efficient on the desktop and server since the Core series. Until the Core came out, AMD had the energy efficiency crown quite solidly on the desktop and server with the K8. Mobile is a different story, the Pentium-M ruled that land with an iron fist, and the Core is the hereditary successor to that. (Although, the Pentium 4m is yet another different story...) The K8 architecture is now old, but it still holds its own quite solidly against Intel's much more recent designs on the server side. I pick my processor based on what's available when I buy a computer. Two years ago I bought a dual Opteron system, because it was the best available. Last year I bought an Athlon64 X2 because it was, quite literally, the best available. This year I bought a Core 2 Duo because it was the best available. I'm hoping for something really interesting for next year's machine, a minimum of 8 cores in a maximum of two sockets is what I'm thinking. In any case, there's really no point in supporting AMD or Intel, just buy whatever is best when you buy a computer. Of course, if one or the other were to fail...that would well and truly suck for everyone.

Re:what marketroids didn't tell us ..

[dbIII]

I'm hoping for something really interesting for next year's machine, a minimum of 8 cores in a maximum of two sockets is what I'm thinking

Yesterday I booted up something like that with knoppix - it was a joy to see 8 penguins on boot. You can get two intel based 8 CPU machines in 1U now (two sockets, 4 cores, power supply in the middle and then the other board with 8 cores) - it will be good to see AMDs four core options since their dual core opterons on 8 sockets is already impressive if on a much bigger board.

What's really relevant ...

[Skapare]

What's really relevant to me is the performance per dollar ... not just dollar of CPU cost, but also dollar of whole system cost (including software, if that goes above zero), and dollar of energy cost (including the cost of shoving waste energy out the back door in seasons I does me no good to keep it indoors).

Re:What's really relevant ...

[SilentOneNCW]

Nice try, but you're off by a great deal.

Re:What's really relevant ...

[raddan]

Why? That observation seems spot-on to me. I recently bought an AMD system instead of an Intel one due to price. I've been primarily a Mac user for the past 18 years, and I needed to replace my aging G4, so I had no opinion either way as far as brand went. I chose my performance level based on benchmarks at Tom's Hardware and other places, and then I sought out to build a machine that met those minimum specs. While it looked like I could get a faster Intel machine if money were no object, the AMD's price point at my desired performance level won by a landslide. This is taking into consideration CPU, motherboard, and RAM cost. The AMD RAM/motherboard/CPU combo costed less than just the comparable Intel CPU.

Of course, if you're after pure performance, then Intel wins, and if you're after that then money is probably no object to you. I think I'm probably indicative of your typical home-machine builder in that I have a budget. Performance per dollar really was the most important thing.

Re:What's really relevant ...

[Anonymous Coward]

Maybe he meant your sig? =P

Re:What's really relevant ...

[raddan]

Well, in that case, he really is off by a great deal, depending on how important that number is to you :^) Didn't even notice the sig.

Real life tasks

[ceka]

It would be more relevant to know how does it perform real life tasks, eg kernel compilation time comparison...

Re:Real life tasks

[Slashcrap]

It would be more relevant to know how does it perform real life tasks, eg kernel compilation time comparison...

Interesting definition of "real life tasks" you have there.

For the majority of the computing population, I would suggest that "real life tasks" would be more accurately defined as downloading and playing porn, rendering MySpace pages and running Norton Antivirus together with the 28 different systray applets installed by Dell during the manufacture of their shitwreck of a PC. Furthermore I would also suggest that more cycles have been burned running just one of the 8735 variants of CoolWebSearch shitware than will ever be used for kernel compilation.

Wow, they really put the squeeze on the high end..

[RalphBNumbers]

I'm all for heated competition, and it's great that AMD can claim integer performance supremacy on the high end again for a while. But at what cost do they make that claim?

The article mentions that the 8222 SE is priced at $2149. So if I want a system with more than 4 cores, I'm bound to pay ~2.5x as much per processor.

I can get a workstation with 8 3Ghz Clovertown Xenon cores from Apple for just under $4000, 8 Opteron cores at the same clock will cost me more than twice that for the processors alone, never mind a motherboard and system to house them.

I'm well aware of the increasing advantages of AMD's bus topology for >4 cores, but with pricing like that who can afford a system to really take advantage of it?

Re:amd quad-cores are true quad-cores with a bette

[Anonymous Coward]

with intel you have to use the chip set for one cpu to talk to another one.

Flat out incorrect. CPUs are fully capable of "talking" to one another over the FSB without any involvment of the chipset.

With current FSB speeds, the distinction between quad-core and "true" quad-core is only relevant to fanboys. By the time it would become relevant, the FSB will be a thing of the past.

Re:amd quad-cores are true quad-cores with a bette

[logicnazi]

This post is very confused.

First of all you can't put two dies on the same cpu, or at least it would be a horribly bad idea. You can put 2 cpu's on a die. Now I thought AMD already did this but they could just package several chips together and I'm feeling too lazy to look it up.

Anyway, yes for Intel chips they must communicate over the FSB. However, as I've recently been finding out they don't do that much communicating. For instance most cache state info is generated just by listening on the FSB. Though sometimes one CPU needs to invalidate a read.

However, not having an FSB AMD's chips don't have a set total system bandwidth they 'use up.' Each chip has it's own memory controller and HT lanes. Perhaps putting the chips in the same package will allow AMD to speed up hypertransport or indirect memory lookup but since AMD doesn't use just an FSB it seems they actually have less to gain than intel by putting many cpus on one chip.

Exactly

[CasperIV]

That's exactly it. People for many years bought Intel chips because they said 3ghz on the box while the AMDs said 2.2ghz. What people took a long time to figure out is that Intel was just bumping the cycles up so that it sounded faster, while the AMDs were getting more work done per tick. I always equated it to engines. You can have a 4 cylinder engine that makes 200 horse power at 7,500 rpms and a 8 cylinder engine that makes 200 horse power at 4,800 rpms. Even though one may have nearly double the rpms they do the same amount of work over time (the equation for horse power). If I increase the rpms (overclock) both, the one doing more work per cycle will have better gains. It is about time that the processor races heated up again, maybe now we will get some real performance gains from the chips. I wonder who will be the first to product another chip that worked as much magic as the Thoroughbred B core in the over clocking world.

Please install this security update for MAC :-)

[Anonymous Coward]

So they found a hole in Safari and exploited it. So one should install this security update?

CPU Speed? Who cares?!

[FFFish]

Sorry, AMD, but I don't get my panties in a bunch over CPU speed any more. The CPU isn't the bottleneck that it once was.

Truthfully, I have not seen a significant benefit to higher CPU speeds since circa 300MHz days. Except for gaming, things seem to always work about the same speed. The rate at which I can type this message is limited not by CPU, but by my fingers; the speed with which I browse the web is limited not by CPU, but by my ability to skim for content; the speed with which I get real paying work done remains about the same.

And even for intensive processing, CPU speed is often less limiting than GPU, HD, or RAM. Doubling my laptop's memory more than doubled its speed; doubling the CPU speed would accomplish bugger-all.

(I lied: one thing that did improve with CPU speed is the performance of Natural Painter. That little puppy loves the CPU cycles!)

Re:Wow, they really put the squeeze on the high en

[Anonymous Coward]

The MAC is just a dualy with two quad core processors. The 4 way 8000 opterons with quad core inter CPU support are a very different thing.

How does it compare with Penryn?

[guidryp]

Wasn't intel recently showing some Penryn benchmarks with up to 50% improvements depending on application.

All pointless till we have a 3rd party compare Penryn to Barcelona. I imagine neither will have much impact till 2008 as both will be production limited this year.

Also AMD need to stop talking and start showing.

Re:Wow, they really put the squeeze on the high en

[dave_boo]

I suppose that when Barcelonas come out, we'll have some affordable octal core loving from AMD. It wasn't all that long ago that we'd have had to go with a quad socket intel mobo to get the performance that is now available from those quad Xenons.

BTW, I've always wondered what a really loaded 3,0 GHZ quad core Clovertown loaded up with as much FBDIMMs as possible would produce in the way of heat.

I currently don't need more speed.

[LWATCDR]

Really I find my current PC fast enough. What I want is lower power and heat for the entire system.
Now if AMD can produce a cheap and silent system with good graphics performance I am all for it. Say something as fast as an X24400 and an Nividia 7600 GT all for about $300 then you have a winner. You will sell millions.
A quad core system? I just don't need it yet.

Shows what you know!

[Anonymous Coward]

The next version of Microsoft Word will require at least 4 of these guys! Just think how much bloat you need to add to slow a processor like that down?!? It will make the bloat of Office 2007 look trivial!

Dedicate 50% of that performance to security

[gelfling]

If there was a way to lock in 50% of that performance for encryption and malware scanning and all the other security gorp that's killing us, that would be great.

Please don't dump it into another golly geewhizbang video or multimedia processor subfunction on the chip.

At this rate, We'll see Penryns before Barcelona

[Glasswire]

Since AMD keeps pushing the ship date of Barcelona out (now Q3) and Intel keeps pulling ship date of Penryn-gen quads (Harpertown) in (now Q4, maybe Q3?) the relivant comparison is not going to be between Barcelona and Clovertown, but Barcelona and Harpertown.
I suppose since, AMD only wants to compare same-clock chips (probably because they won't get higher clock Bars for a while) they may start argueing that we should only be allowed to compare Intel's old 65nm products (not the 45nm) to 65nm Barcelona, too.
And Intel will ship Penryn processors IN VOLUME, so they'll be something you can buy, not just read about on test reviews designed to give the rest of AMDs product line a halo effect from one excellent part.

Combine the two.

[zippthorne]

Bartertown.

Two enter, One Leaves!

Re:Wow, they really put the squeeze on the high en

[Anonymous Coward]

I think you're confusing cores with sockets. The 2xxx series, which is priced very similar to intel, is for use in dual socket systems. Once the barcelona comes out with quad core, you will be able to have an 8 CORE AMD system just by buying 2 of these processors (~$2000 for the CPUs). If you're trying to talk about the 8 8xxs processors, you have to realize that if you populate a motherboard with 8 of these upcoming quad-core CPUs, you would end up with 32 cores, but you're trying to compare this to an 8 core system from intel. Way to compare apples to ambrosia there buddy.

Catch a wave.

[Mockylock]

Regardless of who is faster or has the bigger dick in the past, present or future... There will always be someone ahead of the other, and someone always trying to catch up. Whether you favor one or the other is insignificant against the fact that we're ALL going to benefit from this. After the 2 year dullness in competition pre-conroe, we've actually got something to look forward to.

I've always used Intel religiously. I tried a couple AMD's and they burnt up. Granted, they were earlier chips, but I stuck with Intel after that, even though they may have not been as fast as the AMD. Right now, I know that AMD will come up with something better than Intel.. BUT, I don't mind...

Why?

Price drops and performance gains bruthas. We ALL deserve these long awaited competitions. It's been a long time since we could actually sit back and think, "I've got a CPU/GPU combo that no game that currently made, could actually slow down." And we're doing it at a decent price.

Of course, in time we'll be having heavier apps and using up all the power that it has to offer... but, it's life. It's human nature to spend whatever is in our pocket... whether it be money, or power. Why own a sports car if you're not going to drive it like it's stolen?

Where would you want to live?

[Anonymous Coward]

Barcelona or Clovertown?

Re:amd quad-cores are true quad-cores with a bette

[Gr8Apes]

Basically, AMD gains 2 things by packaging multiple cores as in X2: marketing and cheaper multi-core systems.

With Barcelona they'll actually gain performance improvements with the addition of the shared L3 cache across all 4 cores.

Apple Backwash

[Darth Cider]

When Apple chose Intel chips, a huge mindshare of tech enthusiasts became enamored of Intel. I look into myself and see the irrational plus sign in front of Intel and minus sign in front of AMD, though I'd been a huge fan of AMD before. What the f? Both companies are just awesome innovators. I hope they keep sharpening each other. proverbially, as steel against steel. Gotta love the geniuses in our midst.

Re:amd quad-cores are true quad-cores with a bette

[Anonymous Coward]

Ha! "put 2 dies on the same cpu" gets modded "Insightful." Good job, mods.

Where would you want to live?

[Anonymous Coward]

Barcelona or Clovertown?

Where would you want to live?

[Anonymous Coward]

Barcelona or Clovertown?

Good ! (SC0RE: 5 Insightful)

[Anonymous Coward]

Hiiiii-Oooooooo!

Riiiit-Oooooooo!

Biiiiii-Ooooooo!

Onceagain, Intel got their ass kicked for slacking

[Anonymous Coward]

Intel drops prices like no tomorrow, reschedules core2 redesigns by quarter years. And my guess is they're not doing this just for being generous.

OOOP

[crhylove]

Dude, you haven't switched to Open Office yet?

OOOP = Obligatory Open Office Plug

What socket?

[Unnamed Chickenheart]

What I want to know is - will this new AMD chip use AM2 socket, or a new version?

Re:What socket?

[Lehk228]

i think the new ones are socket F

Re:What socket?

[Unnamed Chickenheart]

This: http://en.wikipedia.org/wiki/Socket_AM2

Claims that there will be two new sockets: AM2+ and AM3.

And at least two versions of the new processors: one will work on AM2 and AM2+, the other one will work on AM3 and AM2+

Re:amd quad-cores are true quad-cores with a bette

[afidel]

One big advantage is most software is licensed per socket so more cores/socket=less $ for software. I know we will be able to get a big jump in performance by replacing the processors in our DL585 G2's with these new quad core when they are released at a low relative to the purchase price for Oracle and 32GB of ram, etc.

Re:Wow, they really put the squeeze on the high en

[Anonymous Coward]

Intel's Core2 architecture is suboptimal and doesn't scale above two cores. Due to the bottleneck inefficiency you lose almost one core when running quad-core. With eight cores it gets even worse.

http://blogs.sun.com/bmseer/entry/woodcrest_scalin g_problems%3F_(fluent_part1)

Where would you want to live?

[Anonymous Coward]

Barcelona or Clovertown?

So many strong opinions...

[stephentyrone]

...so few by anyone who has the slightest idea what they're talking about. Seriously, if you have strong feelings about this issue, and you're not a kernel / high-performance computing / game engine / professional audio/video developer, then you need to get out more. This is just painful to read.

Re:At this rate, We'll see Penryns before Barcelon

[shaitand]

'AMD only wants to compare same-clock chips'

That seems reasonable enough. Unless you are buying the absolute fastest chip on the shelf with a massive premium you want to know which chip offers a superior architecture. Comparing the performance of two chips at the same clock tells you that. Especially since AMD chips are much lower in price than Intel chips. Assuming you don't have a limitless budget you can definately get a rock solid AMD chip that outperforms the equally priced Intel chip at any price point but the very top. Even most of the gaming crowd will be spending $1000 on their system, not on their processor.

'they may start argueing that we should only be allowed to compare Intel's old 65nm products (not the 45nm) to 65nm Barcelona, too.'

I doubt it but it is definately something to keep in mind. AMD runs neck and neck with Intel on performance. AMD always has that card to play, simply but upgrading their fabs the same design that competes with Intel while using an inferior fab process will suddenly blow Intel away. I find it very interesting that AMD's design is so far superior to Intels that their technology can compete with such a huge fab disadvantage. The newer fabs are probably why intel chips are so expensive. AMD probably makes more per chip while selling at a lower price.

'And Intel will ship Penryn processors IN VOLUME, so they'll be something you can buy, not just read about on test reviews designed to give the rest of AMDs product line a halo effect from one excellent part.'

Okay, APPLE might have a problem getting enough chips. Dell might have a problem. HP could have a problem. I've never heard of an individual being unable to get a currently manufactured AMD processor. You can probably buy all the currently manufactured chips from newegg right this second.

Re:At this rate, We'll see Penryns before Barcelon

[ben there...]

Unless you are buying the absolute fastest chip on the shelf with a massive premium you want to know which chip offers a superior architecture. Comparing the performance of two chips at the same clock tells you that. Especially since AMD chips are much lower in price than Intel chips. Assuming you don't have a limitless budget you can definately get a rock solid AMD chip that outperforms the equally priced Intel chip at any price point but the very top.

Before AMD's price cuts, that statement would have been laughable. E6600 at $330 and E6300 at $185 were blowing away AMD chips priced hundreds more. I seem to remember E6600 beating a $900 AMD CPU then. Now it is much closer (and much cheaper), considering the 6000+ and E6600 are both ~$235, but I just looked up AnandTech's review of the 6000+ earlier today, and it is beat by the E6600 by only 5-10% in most benchmarks. Much closer, until you overclock the Core 2 Duo by 25-100% on air, depending on model. You won't see those kind of overclocks from the AMD.

I would consider $235 about right for most gamers. You are probably right about AMD having the advantage for chips less expensive than the E4300 at $125. That CPU is the one that gets 100% overclock, aka "the new budget king of overclocking" according to AnandTech. But I don't know: how many people spend less than $125 on a CPU?

AMD has the wrong target to beat

[steelcobra]

They should be aiming to beat Penryn. But maybe it's just that this is Intel's turn in the lead, like AMD had for the last several years.

Re:Exactly what comprises leadership? Thoughts...

[Douglas Goodall]

I don't think that anyone that's truly interested in the technical aspects of CPUs thinks Intel is a leader in this decade.

I have always been and continue to be primarily interested in the technical aspects of CPUs. In some respects you could say I am an Intel fanboy because I have unwavering confidence in their long term vision and ability to continue the advancement of microprocessor technology. I have primarily used Intel chips since the 8080 and while the Motorola chips had nice general purpose registers, I held on and waited for Intel to evolve the way I wanted.

Starting with the 80286, Intel provided the hardware required to build secure, reliable microcomputers with solid operating systems. The fact that operating system vendors decided not to use the chip features that would have protected the software was an obvious shame. Leadership in the CPU industry is about more than clock rate, or even throughput. A clear vision of the problem domain, the programming languages that would best model and solve the problems, and the best price/performance hardware offerings to integrators and OEMs is what counts in my mind.

When people think about computers, I believe they visualize desktop machines, most often running Windows. Desktop computers are only a small segment of the potential uses of microprocessors to make the world a better place. Windows unfortunately, is a rather poor operating system whose most valuable feature is it's abilty to run with a lot of different PC hardware. When I think about computers, I visualize microprocessors that monitor real-time conditions, and provide some form of useful output, controlling a device, or sending information to another system somewhere. The advent of cheaper hardware and free software has allowed us to posses powerful computer systems running derivatives of Unix (The Best OS ever). We have come to a place where there is enough hardware and software around to solve problems quickly that would have kept mathemeticians busy for decades or centuries.

The challenge now is not running games a little faster, or Windows apps a little faster. The challenge now is figuring ot how to apply the technology such that it benefits us all. Certain application hold great promise to make an actual difference in the real world such as automotive microprocessors that help tune the engines and reduce emmissions. Benchmarks don't make the world a better place. Benefits of the application of technology do though. There is a place for the absolutely fastest processors in critical servers, high end scientific workstations... I am glad there is competition in all areas if microprocessor technology, but I wish we could keep focus on the "golden ring" of making the world a better place.

Re:At this rate, We'll see Penryns before Barcelon

[Glasswire]

Why is it that manufacturing technology is not a engineering advantage in the same way as logic design? AMD is a year behind Intel on process technology (just mainstreaming 65nm now, not doing 45nm until later '08 at which point Intel will have been shipping 45nm for about a year). This isn't some kind of magic, unfair advantage that the semiconductor fairy bestowed on Intel and not AMD. It's a result of prudent choices about where to spend you development budgets and Intel has always cared more about keeping Moore's Law alive than anybody else, and generally it's been to Intel's benefit. So, to seriously suggest that we can't compare 45nm Intel product to 65nm AMD product - OR that we can only compare equivalent clocks. How about we only compare like memory architectures. Oh, no, embedded memory controllers are an advantage for AMD that Intel does'nt have! You see how ridiculous this gets...

AMD always has that card to play, simply but upgrading their fabs the same design that competes with Intel while using an inferior fab process will suddenly blow Intel away.
That's silly. You're implying that for some reason AMD is choosing not use the best process technology it can right now? I think if AMD were in possession of that card, they would have played it. A recurring theme I see among AMD advocates is this persistent assumption that AMD's process gap is somehow momentary, and will soon be erased with AMD having the same process as Intel in the same time frame. This has almost never been true and I think you could successfully argue that the gap is WIDENING to more than a year. With Intel's new tick-tock model, you get 45nm in '07, 35nm in '09 and so forth - I haven't seen any rational observer suggest that AMD could even keep up to that with a 1-year gap, let alone catch up. The more likely outcome is that AMD keeps falling further behind on process.

AMD probably makes more per chip while selling at a lower price.
Well you're first point is absolutely false. Check out the last earnings report, AMD's margins ARE much lower than Intel's - but you are probably right about AMD haveing a lower ASP (average selling price). Which is a huge problem for AMD, not to their credit. AMD would dearly love to raise prices and get their margins back into shape.

Re:Exactly what comprises leadership? Thoughts...

[Gr8Apes]

I have always been and continue to be primarily interested in the technical aspects of CPUs. In some respects you could say I am an Intel fanboy because I have unwavering confidence in their long term vision and ability to continue the advancement of microprocessor technology. I have primarily used Intel chips since the 8080 and while the Motorola chips had nice general purpose registers, I held on and waited for Intel to evolve the way I wanted. Then you and I are of differing opinions. I do not think Intel has a good vision of the future, nor even of the present. C2D was a reactionary CPU developed from in house technology solely in response to AMD's Opteron/FX line of CPUs that were just killing them. Itanium, their idea of the future, appears largely dead, with an ever declining marketshare. The DEC Alpha chip would have been a better bet. Their monopoly pricing practices are now out in the open and are already having a negative effect. Their current price war intended at the minimum to drain AMD to stop innovation is in full swing, but will hopefully be too little too late. Even Intel can't bleed forever, especially if their marketshare shrinks by as much as AMD hopes it will.

Re:At this rate, We'll see Penryns before Barcelon

[shaitand]

'That's silly. You're implying that for some reason AMD is choosing not use the best process technology it can right now?'

I don't think either one is using the best technology they can right now. Quite frankly I think most high end technology companies are on a phased release cycle with developed technology far ahead of what they have announced or are shipping. Hence why AMD can magically pull out massive efficiency improvements as an immediate response to a bad quarter. AMD hasn't upgraded the fab process because they haven't needed to do so in order to have the best released technology. AMD chips have blown away Intel offerings since the first athlon chips were released, this is the first time they have lost that advantage since.

'Intel has always cared more about keeping Moore's Law alive than anybody else, and generally it's been to Intel's benefit.'

In what way? Intel has been living on nothing but brand recognition for years.

'let alone catch up'

It isn't as if they have to develop anything. The technology exists, its only a matter of buying it. If AMD chips can remain competative at 65nm then there is no reason they couldn't skip 45nm equipment and buy 35nm in '09. Just as the smart consumer doesn't buy every technology increment unless they actually need them AMD is wise to avoid wasting funds on new fab equipment with their current process is able to result in chips that outperform Intel chips.

Re:Exactly what comprises leadership? Thoughts...

[Douglas Goodall]

I respect your opion, and the eloquent way you stated it. I appreciate the polite way you responded to me. My feelings about Intel revolve around my being a low-level programmer with a lot of experience in embedded systems and driver writing. I continued to write assembly language long after others had moved to higher level languages. I have a talent for visualizing chips, their modes and registers. I have written a few debuggers. One of the things I like about Intel is that they manufacture a wide variety of processors, supports chips, and peripherals. I even went through a brief phases as a hardware designer during which I created several working systems using a variety of Intel chips. For a long time I had a wire-wrap gun in one hand and a logic probe in the other.

I haven't spent much time keeping up with the specifics of Intel's marketing techniques or pricing policies. I do remember they have had some problems over the years about memory pricing, but that was several decades ago I thought. I have spent a lot of time thinking about RISC vs. CISC and have used both extensively. Over the course of time, my decision to stay primarily with Intel as opposed to Motorola has worked out pretty well, and Apple's recent move to Intel CPUs was a welcome change for me as I like Apple and Intel, but not Power PC programming.

Leadership for me includes aspects such as architecture wide component compatibility. Longevity of skillset viability. For the most part, x86 code has remained fairly consistent and backward compatible. We knew in our hearts to be careful what we loaded into segment registers and should not have been surprised when they became more like selectors. I have never been sorry about time I spent learning about Intel parts. Most of what I learned is still valuable.

Re:Exactly what comprises leadership? Thoughts...

[Gr8Apes]

I don't know if it was eloquent, but thanks for thinking so!

As for Intel, I'm aware they do more than just CPUs. I'm just targeting their CPU side of the business, because that's the one that's causing the grief, as I see it.

Their steadfast devotion to x86 has nothing to do with their "vision", unless that includes IBM's insistence on the cross-license that was required to get the x86 chip into the IBM PC in the first place.

Re: x86 Family compatibility (not regarding IBM)

[Douglas Goodall]

Since the 8086/8088, each successive processor has retained a high degree of compatibility, while enhancements continued to enable advancements in overall capability. The 80186, while not a geat leap, contained a few more instruction modes to make code more efficient. The 80286, while having a protected mode including the task state segment and ldt/gdt addressing more then a megabyte of ram, retained a real-mode capability. Starting with the 80386, the virtual mode allowed a combination of real-mode like operation enhanced with advanced memory mapping and i/o mapping that allowed virtualization of devices. 80486 did all of that and contained a math co-processor. At each stage, the previous capabilities were maintained and the size of the pre-fetch queue was increased thereby increasing the performance of the execution unit. After that, 586/686 class processors continued to arrive with further hardware enhancements such as multimedia extensions. In my opinion, they did an excellent job of retaining compatibility from one processor to the next, and although subsequent processors required specific setups at boot time provided by the bios, there has never been a break in the compatibility chain between 8088 and now. The processors are a thousand times faster, the memory is a thousand times larger, and significantly wider, but the original software still boots ad runs.

I am not quite sure what IBM cross-licensing has to do with that. Please explain as I must have forgotten the implications.