AMD Shows Upcoming Phenom II CPU At 6.0 GHz+

Vigile writes "Today during a press briefing at AMD's offices in Austin, TX the company showed off some upcoming technology that should be available sometime early in 2009. What was most impressive was the overclocked speeds of the pending Phenom II X4 45nm processors. On air cooling AMD showed the quad-core CPU running at nearly 4.0 GHz while with much more extreme liquid nitrogen cooling help the same CPU reached over 6.0 GHz! It looks like AMD's newest processor might finally once again compete with the best from Intel, including its recent Core i7 CPUs."

Comments

first question..

[Hillview]

what's the power rating for this thing at 4 ghz? 250 watts?

Re:first question..

[negRo_slim]

What an odd question to ask when you are so far off the mark.

Re:first question..

[Anonymous Coward]

It's over 9000!!

Re:first question..

[Anonymous Coward]

what's the power rating for this thing at 4 ghz? 250 watts?

If you actually read the article you would have seen "AMD could theoretically have a 3.4 to 3.6 GHz processor at moderate TDP levels (think 125 watts)"

Re:first question..

[Hillview]

Sorry, I guess I was looking at the standard ratings for existing processors from newegg. You know, 2.3ghz phenom at 95 watts, 2.4ghz at 125 watts... You linked it yourself: 65, 95, 125 and 140 Watt

Re:first question..

[Warll]

That link only includes information regarding the current it does have a quick paragraph on the new ones but no TDP is mentioned.

Re:first question..

[compro01]

That's for the current generation Phenoms. You likely want this article, which covers the Phenom 2 procs.

TDP spec at 3.0ghz is 125W, so don't think he's exaggerating that much. I'd guesstimate 150-200W at 4ghz.

Re:first question..

[MaxwellEdison]

Theoretically, he may have been wondering about 4GHz. He, theoretically, may have wanted the power rating of the one they actually demonstrated. Of course he may have theoretically read the article too...in theory that is.

Re:first question..

[Chris Burke]

That link only includes information regarding the current it does have a quick paragraph on the new ones but no TDP is mentioned.

They're almost certainly going to keep the same TDP bands as they have today. Of course the 4GHz and especially 6GHz demonstration are probably well outside those bands.

Re:first question..

[afidel]

Nah, a 33% overclock usually only needs a 10-20% overvoltage so I would say 150W is the top end for the TDP range.

Re:first question..

[ckblackm]

I think a better first question would be: Is AMD still going to be around long enough for this to matter? Disclaimer: I'm a shareholder.

Re:first question..

[Hatta]

1.21 gigawatts

Re:first question..

[khellendros1984]

Note: with that mantissa, the unit is spelled (or at least pronounced) jiggawatts =)

Re:first question..

[compro01]

Voltage sure, but what about the current?

Re:first question..

[Dolda2000]

Power usage grows as the square of the voltage.

For purely resistive loads, yes. I can't say I know the details of how CPUs work internally with regards to power consumption, but I would not be surprised if it is also proportional to the clock speed. In that case, it would be another 1/3 to that, making 192% normal TDP, giving 240 W.

Re:first question..

[compro01]

yeah, CMOS is pretty heavily affected by clock speed. Higher clock means more source/sink current needed to toggle fast enough. And multiply a few nanoamps by a few hundred million transistors and you start talking about quite a lot of current.

Overclocking BS

[MLopat]

This is far from impressive. Showing the overclocking results, especially on liquid nitrogen, is not a good indication of the day to day performance of the processor.

For example, here is a video from 2006 where a Pentium 4 processor is overclocked to 5 GHz.

So no, it doesn't look like "AMD's newest processor might finally once again compete with the best from Intel."

Re:Overclocking BS

[AioKits]

But I get to keep the liquid nitrogen if this is the case, yes?

Re:Overclocking BS

[cbreaker]

Yes, but you fail to note that the P4 was designed to reach insane clock speeds (which it never ended up being able to do.)

You can't get a Core 2 CPU to run at 5Ghz no matter how hard you try.

What this proves for AMD's CPU is that the architecture is able to handle 6Ghz, and the only problem is heat. Heat is a big problem, sure, but it's delt with every day in all sorts of new and creative ways - but usually just from reducing fab size and lowering voltage.

I personally don't care much anymore about who's CPU is 5% faster than the other. I choose what gives me the best options.. And I really have had excellent results with AMD's processors in the past. I have a few Core 2 based machines and they're nice too, no doubt. It just doesn't really matter anymore.

Re:Overclocking BS

[negRo_slim]

On air cooling AMD showed the quad-core CPU running at nearly 4.0 GHz

A highly efficient processor running at clock speeds not seen in standard pc kit since the Pentium 4 era... Sounds like they are regaining their footing to me!

Re:Overclocking BS

[ShadowRangerRIT]

You can't get a Core 2 CPU to run at 5Ghz no matter how hard you try.

Given that the Nehalem is reaching the same speeds or higher on air-cooling, I wouldn't be surprised if Intel could match 6 GHz under liquid nitrogen cooled conditions.

Re:Overclocking BS

[chrysrobyn]

Look, I don't think overclocking in liquid nitrogen is cause for a slam dunk conclusion that AMD is now competitive with Intel, but stating that it's not impressive and not an indication of the performance of the processor indicates a complete lack of understanding of electrical design.

This wouldn't have worked, for example, with the original PPC 7400 (G4) past 500MHz. As it turned out, there was a hard stop getting past that. Finding FMax (maximum frequency) independent of reliability and power concerns highlights design weaknesses. If they can overclock by 50% with adequate cooling, one can conclude they don't have any early or late mode problems preventing higher frequencies, and that metal isn't the limiting factor. In fact, they can easily conclude that the electrical design is sound and that their limit will be what they can qualify from a reliability perspective.

Re:Overclocking BS

[Bobfrankly1]

This is far from impressive. Showing the overclocking results, especially on liquid nitrogen, is not a good indication of the day to day performance of the processor.

If the overclocking results were the only thing in the full article, your argument would be valid. However, your comment indicates that you read the short summary, did a quick search for your P4 overclocking link, and posted for quick mod points from Intel fanboys.

TFA shows the processor benchmarking at 3GHz, and 4GHz with air cooling, likely a custom air setup that would not be uncommon for many self builders. Check Intel's speeds, I'll even give you a link to a vendor. I even filtered for the highest GHz. They are about the same.

So Yes, it does look like ""AMD's newest processor might finally once again compete with the best from Intel." Maybe it doesn't blow them away, but compete with Intel it does.

(This commenter recognizes that raw GHz is not the end-all and be-all of the final experience, but this is the only concrete number we currently have to argue about)

Re:Overclocking BS

[Chris Burke]

What this proves for AMD's CPU is that the architecture is able to handle 6Ghz, and the only problem is heat. Heat is a big problem, sure, but it's delt with every day in all sorts of new and creative ways - but usually just from reducing fab size and lowering voltage.

Well, yes, in the sense that hotter transistors are slower than cooler ones. There's probably no way you could ever run it at 6GHz with air cooling, because even if you could actually dump all the heat produced you'd never get the chip cooler than ambient room temperature. Liquid nitrogen gets the chip well below room temperature so it can run faster.

Lowering the voltage would reduce heat output, but also reduces the speed of transistors. Sure maybe 6GHz may be possible air-cooled with the next generation of fab tech, but that's not really relevant to the current product, and is kinda trivial to say anyway as new fab tech is responsible for us being able to have GHz processors with hundreds of millions of transistors in the first place.

Re:Overclocking BS

[DurendalMac]

OVERCLOCKED on air to reach 4ghz. I've seen Core 2s do the same thing.

Re:Overclocking BS

[Fred_A]

Look, I don't think overclocking in liquid nitrogen is cause for a slam dunk conclusion that AMD is now competitive with Intel, but stating that it's not impressive and not an indication of the performance of the processor indicates a complete lack of understanding of electrical design.

I still want to wait to see how the intel performs when cooled with liquid helium.
(not that I care in the slightest what brand of CPU is in my machine as long as it's current in terms of performance, cheap and not a power hog...)

Re:Overclocking BS

[Bob-taro]

Given that the Nehalem is reaching the same speeds or higher on air-cooling, I wouldn't be surprised if Intel could match 6 GHz under liquid nitrogen cooled conditions.

Here is an example of Core i7 at 5.2 GHz on LN2

Re:Overclocking BS

[Yvan256]

You can't get a Core 2 CPU to run at 5Ghz no matter how hard you try.

I'm sorry, but it's quite easy. All you need is the following:
- Intel Core 2 Duo SP9300 (x 3)
- Duct Tape

Bam! 6.78 Ghz. Done.

Re:Overclocking BS

[Jeff DeMaagd]

It's nice, but I don't think the ability to overclock with extreme measures necessarily indicates that it is better at handling more conventional tasks in benign conditions.

Re:Overclocking BS

[Anarke_Incarnate]

You must have mistaken frequency for speed. The clock is not the speed of the processor at its tasks, but rather the frequency of the processor. Let the benchmarks speak for themselves.

Re:Overclocking BS

[IorDMUX]

personally don't care much anymore about who's CPU is 5% faster than the other. I choose what gives me the best options.. And I really have had excellent results with AMD's processors in the past. I have a few Core 2 based machines and they're nice too, no doubt. It just doesn't really matter anymore.

When they're about equal, I choose AMD, so that next time I build a computer I'll still have a choice.

Re:Overclocking BS

[cbreaker]

Absolutely. I agree with you.

Heck, even if the Intel solution might be a LITTLE better, I'll still pick AMD a lot of times.

And, at work I have the say as to what systems our clients will order for VMware hosts, and I always pick Opterons, even if they're a little more expensive. With AMD you get better multi-core performance when you're using a lot of RAM - perfect for VMware.

This is likely to change when Intel releases their Server CPU's with what is basically HyperTransport and an embedded memory controller (which they said was a BAD idea...) But I'll still choose AMD because I want to support the company that's moving the technology forward.

When you think about all of the big improvements to the x86 architecture in the last 10 years, it's almost all AMD innovations. A high-speed interconnect bus with no north bridge. An integrated memory controller. x64. The list goes on - AMD led the way, and it's amazing the arrogance you hear from the executives and product managers at Intel. They actually had the balls to say, when questioned about the similarities between Intel's new CPU and Athlon/Opteron, "Smart people can come up with the same ideas." Ohh, sure they can, but these engineers haven't been living in a box for the last 8 years, so that's utter bullshit. Unbelievable.

Re:Overclocking BS

[MLopat]

Did I even use the word speed anywhere in my original post? No. But thanks for restating something that just about everyone here already knows and frequency and speed. Also, reread the post and see that I'm infact agreeing that the frequency should not be an indication of performance.

Re:Overclocking BS

[MLopat]

You're quick to assert that the AMD processor will compete with Intel but then turn around and give a disclaimer about knowing GHz != Performance. So again, my point is that the overclocked speed does not imply AMD is back on Intel's level.

Re:Overclocking BS

[snuf23]

GHz does not equal performance. Which is why a 2GHz Core 2 CPU can beat out a 3.6GHz Pentium 4.

Please don't bring back the myth.

Re:Overclocking BS

[TheReaperD]

I couldn't agree more.

Even all other things being equal, I still tend to prefer AMD as they have a better track record of supporting upgrades without having to change out your CPU, RAM and motherboard every time. Also, if your board burns out, you're far more likely to find a new retail motherboard for AMD to replace it. Intel, a lot of times, your only option is eBay.

Re:Overclocking BS

[ub3r n3u7r4l1st]

There is a hard limit for the electrical circuit design on how fast you can go. Able to reach 6GHz under any circumstances means the maturity of AMD fabs. You can't just use liquid nitrogen and overclock a 80486 to run at 5GHz.

Re:Overclocking BS

[Zephiris]

Funny, people who have done it disagree.

Re:Overclocking BS

[superpulpsicle]

Who would pay $1000 for a processor to keep a processor overclocked at 5Ghz running 24/7? Nobody. People who do this are doing it for benchmarks marketing reasons.

Re:Overclocking BS

[KikassAssassin]

The reason these overclocks are significant is that the Phenom wasn't able to reach the clock speeds that AMD originally expected it to, and that's one of the reasons it's had such a hard time competing with the Core 2. They were expecting to be selling 3 GHz Phenoms soon after launch, but the first Phenoms to ship were only 2.3 GHz, and even now the fastest Phenom chip is a 2.6 GHz part that struggles to overclock much past 3 GHz or so regardless of the cooling that's used. Even though clock speed by itself isn't a good indicator of performance, especially when comparing different architectures, if a processor was designed to be competitive at 3 GHz and they're only able to get it to 2.6 GHz reliably, that's a big problem.

If AMD has solved their clock speed limitation issues with the Phenom 2, that may help put them back on track to being competitive again. It's still too early to start celebrating until we see benchmarks, but it's good news nonetheless.

Re:Overclocking BS

[Bobfrankly1]

Your point as you wrote it, is because they posted an overclocked speed, and you're because you're against quoting (or reading from) the full article, that the chip can't compete. Solely based on the overclock speed. Your statement makes it quite clear:

"So no, it doesn't look like "AMD's newest processor might finally once again compete with the best from Intel."

There are other speeds provided besides the overclock speed. Disregarding these speeds makes you look like an Intel fanboy retard. Try reading the whole article and the whole comment. You might learn something.

Re:Overclocking BS

[MLopat]

Did you read the article? I did. See I have that nifty little * beside my name which means that I pay a couple extra dollars so that I can read the articles before everyone else and leave informed comments.

So to quote from the FULL article "I do not believe that AMD wants to get back the performance crown at this time, namely because they do not have a design that is head and shoulders above what Intel currently has. So they are content to be 2nd best, and keep Intel in sight in terms of performance and features." -- even the person writing the article contends that AMD is not on par but wants to keep intel in sight. Further, if you read the article again, there is no mention of benchmarking or any real world comparison of the AMD chip and the intel chip as you have asserted.

So looks like you're the "fanboy retard" that could learn something from following your own advice. ;)

Re:Overclocking BS

[Bobfrankly1]

What do you think competing is?

To contend emulously; to seek or strive for the same thing, position, or reward for which another is striving; to contend in rivalry, as for a prize or in business; as, tradesmen compete with one another.

The article also implies that AMD may be capable of out-clocking Intel's current CPU line, but doesn't because of the costly race that would follow. (See the paragraph starting with "The big question") And also, from the SAME full article:

For now, we have to sit and wait for the Phenom II to hit the market. Intel certainly needs the competition, and this product might just deliver enough of it to keep things interesting.

I did error in stating that the Phenom II "benchmarks" at those speeds, when I should have said that it "clocks" at those speeds. But I did read the whole article. Next time you read early, read it all instead of trying to get mods points off of a quick post.

Re:Overclocking BS

[ppanon]

Regardless of the processors success, there is a reason Intel is now designing chips that are remarkably similar to the way AMD has been doing it for a while now.

Yes, it's called diminishing returns. Intel initially used large caches to make up for the lower performance of having a separate north bridge, because designing an interprocessor cache coherency mechanism and NUMA support is the tougher job. They could rely on their financial and manufacturing strengths to push fab and process improvements and support cache increases. However Intel have reached the point where the cache is already the biggest portion of the CPU's transistor budget. Scale-down process improvements can't help them as much to increase the cache if they also want to increase the number of cores.

Now Intel have to tackle integrating the north bridge and supporting NUMA at higher clock speeds than AMD dealt with when they started HT. Meanwhile, AMD have already got that part of their design debugged and ramped up, and almost certainly have some really good empirical data on real world performance and the strengths and weaknesses of their design approach that's now on it's 3rd major design iteration. They will be reaping the less risky benefits of increasing cache size as they move to new manufacturing processes with bigger transistor budgets and yet still have smaller transistor budgets than Intel at the same process size, thus reaping greater yields.

I can't help but wonder if Intel didn't expect AMD to stumble on the harder problem of NUMA, and was hoping to kill off AMD before Intel would ever have to tackle NUMA and north-bridge integration. That would have allowed Intel to tackle the problems at a more leisurely pace. It didn't help that they wasted years on the memory fiasco, trying to leverage their market dominance into splitting monopolistic profits with a RAM patent troll. While they've currently got the performance crown, I think Intel's behind the 8-ball now.

I remember when people talked about how the clock speed headroom of the incoming P4 was going to allow them stomp AMD on performance. I countered that I suspected the huge pipelines in the P4 meant that faulty branch predictions and other pipeline stalls would make it very hard for the P4 to come close to its maximum theoretical performance on real-world workloads and that AMD's processors would remain competitive. I think people are ignoring some significant details in a similar way here. Intel is facing a very complex engineering challenge, that's gotten more complex due to their procrastination. AMD's biggest challenge was finding funding for fab upgrades in a credit crunch, but it looks like they've found some Arab angel investors. By splitting off the fab company, they keep the critical IP and design talent and shed a lot of the debt. They've also got one of the best processor design managers ever, Dirk Meyer, running the slimmed down design company. I think they're in very good shape.

Patience

[Anonymous Coward]

If they just wait six months, their new chip won't be such a hot product and thus won't need liquid nitrogen to reach 6GHz

All good but...

[Anonymous Coward]

Will it run Vista?

Re:All good but...

[drachenstern]

Fail:

All Good but... Will it blend?

Now, that's the question! Better yet, can you blend an i7 and a Phenom II?

Misleading title

[ShadowRangerRIT]

If you need liquid nitrogen to boost it to 6 GHz, it's not all that interesting. Nehalem 2.66 GHz offering has also been shown to overclock to 4 GHz on air cooling, and some people have got the 3.2 GHz offering up to 4.5 GHz on air. On GHz they're roughly the same, possibly with a slight Intel edge.

I thought both companies were ditching the GHz war and fighting for actual performance supremacy? What's with the silly "my GHz is bigger than yours" competition? Do we have PPW numbers, or just press releases that mean nothing?

Re:Misleading title

[LWATCDR]

Actually the i7 last time I checked only went to 5.9ghz on LN2. And they showed 4ghz on air so yes this is pretty interesting depending on the price point.
An over clocked $250 AMD Phenom II could compete very well with a $1000 i7.
That would be very interesting.

Re:Misleading title

[BlackSnake112]

I have had my Q6600 running at 4GHz for three weeks 24/7 on air. 4GHz is not that hard. The system didn't crash, everything ran fine. I didn't have to turn the heat on for those three weeks either, but cranking a 2.66 GHz to 4.1 GHz is going to create more heat.

It did take a bit of tweaking to get it there. But I did it, and now it is running back at stock speeds. This machine is a my DVR it doesn't need to run at 4 GHz to record TV. Yes I have a much less power hungry system on order. I was just testing things out with what I already had. The biggest gains I saw at 4 GHz were: boot times were faster, got higher CPU scores on benchmarks (3d mark etc), not needing to turn on the heat for a while.

Re:Misleading title

[Bert64]

The Phenom seem to perform somewhat better under linux, or at least gcc produces better code for them than it does for intel chips...

(note, this is based on 64bit gentoo, gcc 4.3.0 compiled with -O2 and appropriate cpu type setting on a 2.3ghz phenom 9500 and 2.4ghz Q6600)

Re:Misleading title

[LabRat]

"If you need liquid nitrogen to boost it to 6 GHz, it's not all that interesting. Nehalem 2.66 GHz offering has also been shown to overclock to 4 GHz on air cooling, and some people have got the 3.2 GHz offering up to 4.5 GHz on air. On GHz they're roughly the same, possibly with a slight Intel edge."

Show me an Intel chip that doesn't require LN2 to reach 6Ghz, and I'll agree with you that it's not interesting. The fact that they've been able to at least match an intel offering is very interesting to me, personally...given the 2 years of having their asses handed to them by the wakened Intel giant.

"I thought both companies were ditching the GHz war and fighting for actual performance supremacy? What's with the silly "my GHz is bigger than yours" competition? Do we have PPW numbers, or just press releases that mean nothing?"

Well, back in the day of the P4 versus Athlon...you might have had a point since there was such a huge gulf between the two architectures in instructions per cycle (IPC). These days, the IPC is much closer so Ghz is a decent (though not perfect) measuring stick for coarse comparisons. Obviously, the most important metric is a set of benchmarks on the application that YOU are interested in..but as a first cut Ghz will do these days (it's the closest we've had to an apples-to-apples since the old k5/PIII days). Others have already talked about how reaching the "max freq", on air and cryo-cooled is more of a testament to the engineering soundness of the design, and not necessarily indicative of the consumer-level performance to be expected. Again..benches (and the PPW numbers you cited) are where the real action is at, but it's hard to fit a full set of benches on a shelf tag so Ghz still has its place in marketing. For the first time in a while, the Ghz comparisons actually have some meaning, so enjoy the brief window of sanity in number comparisons between the two camps.

Re:Misleading title

[darien]

Even their 9850 Phenom, 2.5Ghz, gets an average of less than half the performance in many benchmarks and many reviews, compared to the classic 2.4Ghz Q6600. The revision stepping that fixed the 'TLB errata' in hardware, rather than the 'slow software workaround' saw...no performance improvement clock for clock?

Can you provide references? I found the TLB fixed Phenoms more than 10% faster than their predecessors, making the X4 9850 convincingly faster than the Q6600.

Re:Misleading title

[basicio]

AMD has consistently bested Intel in performance-per-watt comparisons. Their newest released processors (Shanghai) already offer better performance-per-watt than Intel's pre-i7 processors (see Ars' Shanghai review here). They are behind still, yes. But they are catching up.

Re:Misleading title

[Ant P.]

AMD from the beginning treated 64-bit mode as a first-class citizen.

They also treated GCC as a first-class citizen, while Intel is content with trying to compete with it using its own compiler that generates vendor-sniffing code like a bad geocities website.

Re:Riight...

[bassgoonist]

I didn't know the ground was tethered down to start with.

Only nitrogen?

[cheetham]

Anyone ever tried cooling a CPU with a continuous flow of liquid helium? :)

Re:Only nitrogen?

[Anonymous Coward]

Liquid helium?! Great! Now my processor can talk with a squeeky voice and giggles!

Re:Only nitrogen?

[Etrias]

Meh, not as funny as you would think.

Re:Only nitrogen?

[Cyberax]

Even better - with _superfluid_ liquid helium :)

Re:Only nitrogen?

[CaptainPatent]

That's nothing, I just finished overclocking my server by giving it a continuous feed of liquid hydrogen...

Man that puppy runs cool... I think I'll celebrate with a smoke!

[INSERT CaptainPatent INTO DARWIN_AWARD]

Re:Only nitrogen?

[tylerni7]

Liquid hydrogen is about 16 degrees warmer than helium, and is usually used as a fuel, not a cryogen so I don't know why you would use it over liquid helium...

Oh that was a joke?
I see... carry on.

Re:Only nitrogen?

[fbjon]

I prefer a block of helium ice on top of the processor.

Re:Only nitrogen?

[cerberusss]

[INSERT CaptainPatent INTO DARWIN_AWARD]

INSERT INTO SQL101
(nick)
VALUES
('CaptainPatent');

Yes I'm being pedantic here ;-)

Re:Only nitrogen?

[kirk444]

Yeah, but all my music sounded really high-pitched afterward.

Normal usage, please

[TheDarkMaster]

Is useless a test using overclock in this case, i opened this article tinking the CPU runs 6GHz at stock config!

Re:Is GHz even important?

[lee1026]

GHZ is actually very important. Given that all else remains the same, a 10% increase in clock speed is not greatly different from a 10% improvement in performance in CPU bound applications. Comparing GHZ across different designs is a rather bad idea, but that is not all that is going on here.

Don't try it, you might get sued...

[G3ckoG33k]

www.patentstorm.us/patents/7243507/description.html

Unbelievably, you didn't try http://www.google.com/search?hl=en&q=liquid+helium+cpu+cooling&btnG=Search

Re:Is GHz even important?

[LWATCDR]

Ummm. No...
We learned that clock speed wasn't the end all and be all of performance.
In this case both Intel and AMD are getting good performance per cycle so upping the clock is a good thing.
If it can do 4Ghz on air then yes it is getting in the the competitive range.

Basically

[Sycraft-fu]

GHz is brought up when your chosen platform is on top. If you aren't, it's downplayed.

For example the original "MHz myth" was started by Mac fans. When they first went PPC, Apple had a large lag behind Intel in MHz. Well, the Mac fans were all excited about this new architecture and kept talking about how PPC has a positive second derivative of MHz and x86 had a negative one and so on and so forth. They were all excited about how they'd be ahead in MHz in a few years and basically equated MHz to performance.

Well that didn't come to pass. PPC didn't scale up in MHz fast and x86 did. So all of a sudden they started whining about the "MHz myth" and saying that it didn't matter, performance did. When their platform wasn't going to be on top it changed from important to worthless.

Same shit here. When Intel had the high GHz chips, AMD heads were up on the fact that AMDs did more per clock. Now if AMD has the high GHz chips, they'll be touting that as being the measure of awesomeness.

Me, I'll just keep buying what does the job best, forget the clock speed.

Re:Basically

[thepotoo]

You seem like the right person to ask this of, so, how exactly do I choose what's best for the job?

I want a good, cheap, stable, processor that's going to be able to handle every game made in the next few years (the same thing every home user wants).

I can't really judge by GHz, since my ancient 1.6 GHz processor is enough to handle most modern games when overclocked (2.13 GHz). So what do I look at? L1/L2 cache? FSB? Does the tech (45 nm) factor into speed at all, or is does it just give a general idea of how advanced the chip is?

I want to know more about the underlying technology and how it impacts real-world performance rather than just "buy a Core 2 Quad Yorkfield 2.83GHz 12MB L2 Cache, n00b": if I wanted a recommendation for a specific chip I could use one of a million benchmarks and pick one that's rated highly on Newegg.

I know, I know, JFGI, but I can't find a decent explanation anywhere. Even Wikipedia (while having plenty of great technical info) doesn't really tell me how having (for example) a larger cache will improve performance.

Re:Basically

[Chris Burke]

Well that didn't come to pass. PPC didn't scale up in MHz fast and x86 did. So all of a sudden they started whining about the "MHz myth" and saying that it didn't matter, performance did. When their platform wasn't going to be on top it changed from important to worthless.

Same shit here. When Intel had the high GHz chips, AMD heads were up on the fact that AMDs did more per clock. Now if AMD has the high GHz chips, they'll be touting that as being the measure of awesomeness.

Me, I'll just keep buying what does the job best, forget the clock speed.

Right, that's the best thing to do, because despite the changing opinions the fact is that the Megahertz Myth is and has always been true. It's worth pointing out that at certain points in time, like when AMD and Intel were racing to 1GHz, the architectures being compared were similar enough that MHz was a decent first-order measure of relative performance. That went right out the window with the P4 which was designed for highly marketable MHz numbers at heavy cost to IPC* thus more or less necessitating AMD to play up the MHz Myth. Now even though both companies are back to PPro and K7-derived architectures they're still far enough apart that MHz isn't very meaningful. Nevertheless, MHz is half of the (clock frequency)*(IPC) performance equation, most importantly the half that doesn't change from app to app, so it's still important to demonstrate that processors have frequency headroom. Especially since squeezing out more IPC is difficult and done in 5-10% chunks over the course of a major processor revision.

Anyway, if you really care about performance and not Rah-Rah bragging rights (or whining rights), benchmark scores are what matter. Or better yet, the performance results of whatever you personally find important, though that can be hard to do. Still, that's no reason why AMD or Intel shouldn't pursue higher frequencies, or show off when they can.

* There was an engineering justification for this strategy, they had charts showing how they could continue ramping frequency, even at the cost of having to add more pipe stages, and still have a long-term performance advantage. That never came to pass, because their chart didn't account for increasing leakage current as transistors shrunk. In any event, I guarantee you that the mandate for clock frequency over all else came from management/marketing, and the engineers just had to find a way to make that work.

Re:Basically

[poetmatt]

You are (possibly unintentionally?) turning this into a straw man argument.

It's not the processor that really dictates what you can handle. It's the graphics card moreso.

As long as you have a mobo that can handle PCIEx16 (minimum) and a nice processor, well you're good for quite a while. Meanwhile, anyone buying any motherboard right now is potentially fucked not because of processors or graphics, but because of USB3 coming out. Since no current motherboards will be able to support that speed without a drop-in PCIE card, you can imagine that in a year or two when USB3 is commonplace anyone without it is going to be screwed.

If you want to find unbiased reviews, stick with techreport.com and follow their articles. They are the company that outed other companies (and tested their own as well), to see who was willing to give positive reviews based on being bribed, such as tomshardware.

Re:Basically

[danbert8]

My advice is get any processor you want that is currently sold. Unless you are doing some serious number crunching, very rarely is the processor the bottleneck in the system. Buy a $100 or less processor and throw the savings into RAM and the best graphics card you can afford. Hell even getting a faster hard drive will make a bigger performance difference than a 2ghz processor and a 3ghz processor for 95% of applications.

That's why I suggest AMD systems. Not because they are faster, but because I can get a processor that can handle pretty much anything and then some for $30 with very low power draw and toss in a much lower power demand AMD chipset motherboard and they will save big in the short term on the hardware and save in the long term on power consumption, all while having an indistinguishable difference in performance to a $500+ processor.

Re:Basically

[maxume]

"screwed" in what sense? In the same sense that anybody who buys a PC in the next year or two is going to be when USB4 comes out?

I guess people that have an urgent need for the bandwidth provided by USB3 will have to spend a little more money than people who think the bandwidth will be a nice add-on to the baseline feature set, but your description of the situation is simply hyperbolic.

Re:Basically

[Bob-taro]

They are the company that outed other companies (and tested their own as well), to see who was willing to give positive reviews based on being bribed, such as tomshardware.

Really? Do you have a link to that story? I've read THG for years, so I'd like to hear the full details of that claim.

Re:Basically

[Sycraft-fu]

For games, doesn't much matter. Get a dual core chip that's reasonable and you should be fine. Games do use the CPU but the GPU is by far more important. You can get a quad if you really want but at this time very few can use it at all, and those that can don't tend to be that efficient. A good dual core from the mid range area from either manufacturer should work well.

65 vs 45 nm isn't that important except in terms of energy usage. The 45nm chips are going to use less power for equal performance. However this again isn't a huge deal since the GPU is likely to be the big drain in the system.

Cache isn't all that big a deal. Again, just get whatever the midrange is. Games aren't an area where cache seems to make a large performance difference.

More or less, while these things can make a difference, they don't make enough to justify that much worry or money. You will probably find that a $250 processor works pretty much as good as a $1000 processor, whereas a $300 graphics card is going to be 50% faster than a $150 graphics card. Thus it is clear where your money should go.

Personally I have a Core 2 Duo 2.66GHz 65nm chip and it works just fine on all the games I've thrown at it. In general, when games are limited it isn't the CPU it is either the refresh rate of the monitor (it isn't useful to go above that and as with all LCDs mine is set at 60Hz) or the GPU. Now keep in mind the GPU I have is a GeForce GTX 280. So even a GPU that heavy hitting doesn't really seem to need more CPU, for all the games I've messed with.

If I were to build a gaming system today my strategy would be as such:

--Get a midrange CPU. Something probably not more than $300, but not less than $150. Maybe a quad core since I also do audio work, but I'd be looking more at dual cores. I'd make sure it supports DDR2 RAM, since DDR3 is currently too pricey to justify the small gain.
--Get 4GB of RAM. It's cheap, why not.
--Get a nice big drive since bigger drives are faster and games are not getting any smaller.
--Get a video card such that I can afford to get a new one of the same price once every 12-18 months or so.

That last one is key: Your video card is important to games, and it gets outdated real fast. You can't buy one that won't, because new technology comes out all the time. You can drop $2000 on an insane multi-card setup, and it'll still be outdated soon. So, the right answer is to buy less card, more often. I say make it a yearly target. You aren't necessarily going to buy that often, but that's a good target to make sure your price is realistic and you really don't want to buy more often than that. So whatever you can afford per year, get that. Then, when the next worthwhile upgrade in that price range comes along, get it.

That's what I did. Prior to my 280, I had an 8800. They both cost me about $400. I can afford to spend that every year (I spend a lot on my computer, it's important to me). In that case, it was more like 16-18 months, which is fine. You keep your card until there's a new one worthwhile and/or you find a game that doesn't run well, but you are ready to upgrade yearly. Used to be I couldn't afford so much, so I used more midrange cards. I had a GeForce 3 Ti 200 back in the day. Wasn't top of the line, but I could afford to get it, and then to replace it next year if needed.

So get a midrange CPU, plenty of RAM, and a video card that you can upgrade and you should be fine. CPUs have pretty good life these days. It's videocards that are obselete all the time. Good news is that videocards can be gotten for reasonable prices. For example an ATi 4850 will run you about $150, less after rebate. However it is enough to run any game out there at high detail at a reasonable rez. My bet is it lasts more than a year, but at $150, it isn't unreasonable to replace next year if you need.

Re:Basically

[thepotoo]

Thanks for techreport.com, I'd somehow missed that site.

I'm not sure why exactly I'd be screwed when USB3.0 comes out. Will it break my existing mouse and keyboard? How about my flash drives?

And if I really need USB3 functionality, I'll be able to buy that PCIE card, so I fail to see what the problem is.

I am well aware of the benefits of a good video card, and I opted for a $25 dollar processor (2800) and a $150 VC (7900GS) several years ago, and they served me (really) well. However, they are in need of an upgrade (especially as I play RTS games and a single-core seems to be holding me back), and I'd like to know what to look for in a good gaming processor.

(Sycraft-fu's post further down really answered my question).

Re:Basically

[Sloppy]

how exactly do I choose what's best for the job?

Buy a chip that costs about $100 and then be happy with it, or buy one that costs about $300 and be extremely happy with it.

Re:Basically

[Sj0]

Simple test:

Are they selling it to you? It's probably way more powerful than you'll need for a long time. Go with the cheapest one you can get on the newest architecture you can find.

You said it yourself, your ancient processor overclocked slightly (it does matter whether it's AMD or Intel to know how ancient that is) is almost enough to play today's games. No matter what you buy today it'll be way more powerful than either an Athlon XP 2000+ or a Pentium 4 1.6Ghz, probably 2-4 times before counting the dual-core aspect.

Good processors are getting so fast and so cheap, it's not really even worth figuring out. Go with the one that gives you the most warm fuzzy feeling inside.

Re:Basically

[Jorophose]

The nice thing about current GPU/CPU tech is that a Athlon X2 2.6GHz with an 8800GT/HD4850 is considered baseline by Ars, and can play games at 1680x1050 comfortably. That's pretty incredible.

Re:Basically

[poetmatt]

Umm, no. USB3 will be backwards compatible with 2, so it's not like we won't be able to use USB3 devices on a USB2 computer or vice versa.

The issue is that there is a huge realm of hardware devices that can suddenly reach fruition when they can function at 600MB/s that are simply not available right now.

Example: anything streaming (HD especially). It even opens up a reasonable possibility of an external graphics card for low level tasks. I honestly can't imagine what else as I cannot see the future, but going from approximately 30MB/s cap to 600 (megabytes, not bits)? Oh, also storage. This will be basically external Esata connection speeds for every single port on a pc. It could also mean enough bandwidth to output via USB3 to high def, as in possible DVI replacement (DVI single channel is 4gb/s which is the same as USB3).

This also means a necessity for a huge total motherboard bandwidth increase on par with how much a graphics card needs for each port, so either we're talking a seriously fast USB controller or we're talking a huge increase in bus period. Effectively this could bring about reasons for more necessity for quadcore and above bandwidth devices as well.

USB 3 has been like 4+ years in the making. I don't think we'll be likely to see USB4 soon. This isn't like a "don't buy a videocard this generation, you can wait till the next".

Re:Basically

[poetmatt]

Aww, I apologize. It was the site that works with techreport, dailytech. why do I trust techreport? Because they also tried to bribe their own company secretly in this process as well to ensure honesty. They also have a much more up to date system building guid3e (monthly) with straight newegg links.Here are two of the related links:
http://www.dailytech.com/PR+Firm+Owner+in+Hot+Water+Over+Game+Review+Site+Ownership/article13085.htm

and also

http://www.dailytech.com/Pay+to+Play+Uncovering+Online+Payola/article7510.htm
is the bribe/payola one. Read carefully about tomshardware in there. It's been well known that they are in the payola category and aren't afraid to use outdated drivers in tests or refuse to test outright certain products that would bash certain competitors, such as the HD4870x2.

Re:Basically

[poetmatt]

Oh, no. USB3 will be backwards and forwards compatible. You can run USB3 on USB2 devices and vice versa, but the speed difference is 20x faster so it's a big deal (way faster than firewire). AKA 600MB/s. Meaning that you could tune HD tv and use ridiculously fast flash drives/it could be straight competition to Esata as well.

Gaming processor = look at techreport's system builder guide on the front page. They show justification for why they choose something and newegg links to the whole system and is updated monthly I believe, so it's pretty easy to compare. You'd be amazed how cheap the stuff is sometimes :) In my personal opinion, phenoms are dirt cheap for a quadcore and getting a ati/amd 4850 can get you some monster performance real cheap ($170 for a vcard that very competitive right now that you could drop in another for sli at a later time to run as fast as the baseline fastest card on the market).

Re:Basically

[Michael Hunt]

It won't be 600MB/sec. It won't even be a theoretical 600MB/sec. Assuming that USB3 uses the same 10B/8B coding scheme as earlier incarnations, your peak throughput will be 480MB/sec. In reality, you'll be lucky to hit 300-400, assuming that USB3 scales roughly the same as USB2 does. Of course, crappy cables inducing retransmissions and such are likely to be a far greater problem at a 4800Mbit/sec line rate than they are at 480Mbit/sec, so even that number is only likely to be semi-reachable in the real world.

Re:Basically

[poopdeville]

I wrote this in response to another, and decided against posting it (since the GP specifically asked about gaming, but I missed that), but then decided to use it in response to your post. It expands on your points.

Addressing the GP's question: It depends on your application/use-case. A database server can do fine with a relatively slow multi-core CPU, as long as it has fast enough IO and a LOT of RAM. A scientific computing/numerical analysis machine will likely need the fastest FPUs it can find -- hence the popularity of the PowerPC architecture in that field. Depending on the domain of scientific computing (discrete data mining or other discrete mathematics problems), something like a Sun UltraSPARC T2 could potentially do very well. These sorts of machines often have a LOT of RAM too, since hitting swap can turn a 10 hour long run into a year long run. Gamers need fast FPUs, usually provided by the graphics card.

The average user will probably benefit most from a quick IO subsystem combined with 4GB of RAM. Fast IO will allow your applications to be loaded quickly, and lots of RAM will help keep your applications and data cached, so you won't have to hit the disk for them again. (Which means the second time you start an application will be several orders of magnitude faster) Fast processors are nice, but only if you actually do processing. There's no point in keeping a 4.0 GHz Core Quad Extreme idle, or waiting on disk or user input, except for bragging rights.

I wouldn't worry about things like the L2 cache much as long as you get a modern processor or are doing serious number crunching. A cache is only useful if it gets "hit" -- if the value the processor is looking for is already stored in cache. A cache miss means going to RAM or disk (or reprocessing) for the value. A MB is enough to store the most used parts of your OS's kernel. Scientific computing applications are often designed to fit in the L1 or L2 cache -- if the main loop is going to run a hundred trillion times, it's better to keep its instructions in the cache so you don't have to go to RAM for them, wasting 2-4 cycles each time. Also, if the main loop is bigger than a cache, you will ALWAYS have to go to the next layer of memory, either from L1 to L2, or from L2 to L3 or RAM.

Other processor subsystems are similar -- you would already know what you needed if you "really" needed it.

If you want to compute faster, figure out which component/subsystem is slowing you down. And look for specific recommendations about that component/subsystem. The rest of your system will fall into place around it, at least as fast as your current one, since everything has gotten incrementally faster. The adage regarding premature optimization is true of both software and hardware.

Re:Basically

[thepotoo]

All I can say is thank you. That post gave me fantastic information in an easy to understand format. You should repost it to Wikipedia so that more people can see it. Especially the stuff about L1/L2/L3/RAM/Swap, that's the best explanation I've ever heard.

Re:Basically

[cymen]

Broadcast HD tv maxes out at approximately 19 Megabits (yes, bits not bytes) per second. That is the total bandwidth but many stations have multiple channels sharing that bandwidth.

Re:Basically

[andrewd18]

I've always purchased my CPUs with the following rules of thumb:

1. Pick the CPU slot or form factor that has the longest roadmap. With my last computer, that was the AMD 939-pin platform. This will allow for future upgrades to the CPU, even if 3-4 years have passed.

2. Once the form factor has been chosen, do the research and find the best performance-to-price ratio. If it comes down to a few different processors, pick the one with the highest FSB and/or memory onboard cache.

3. Get a quality cooler for the CPU, like Zalman, that will last a while, and can be cleaned and transplanted onto a future CPU.

Re:Basically

[poetmatt]

Uh, do you have a link to show that for my own understanding? I seem to recall that not only do USB2 devices have a difficult time receiving HD to 1080P but also not all "broadcast HDtv" is transmitted at the same bandwidth either. Not that the difference should be enormous. From what I read, you're referring to digital tv, which is not HD broadcast.

Re:Basically

[Ant P.]

(way faster than firewire). AKA 600MB/s.

Just like USB 2's 480mbps was "faster" than Firewire's 400mbps? No thanks.

Re:Basically

[mollymoo]

I want a good, cheap, stable, processor that's going to be able to handle every game made in the next few years (the same thing every home user wants).

Every home user? Erm, no. For most home users performance rendering web pages, using Office and displaying their photos is much more important than playing the latest 3D FPS.

Re:Bahhh! Legion hardware did that on a i7.

[Anonymous Coward]

The Phenom II will fit in my AM2 motherboard (which started with an Athlon 64 2.0GHz and currently has an Athlon x2 2.6GHz) and use my existing RAM. The intel i7 will not. The intel i7 is significantly more expensive than anything AMD has too.

Re:And ?

[randyest]

Yep. And that means a dual-core version will leave you limbless.

I've stuck with AMD

[rsilvergun]

First, I can put together an AMD box equivalent to an Intel for about $200 bucks cheaper. That money goes into my Video Card and I wind up with a better overall System.

Second, AMD clearly differentiates their product. An XP 6000 is faster than a 5000, etc. Buying an Intel CPU is a chore (and make sure you get the right board, That's not always clear either).Basically I'm Lazy, and Intel's made it a pain to pick the right processor.

Re:I've stuck with AMD

[Tetsujin]

First, I can put together an AMD box equivalent to an Intel for about $200 bucks cheaper. That money goes into my Video Card and I wind up with a better overall System.

I just bought a new computer, looking at quad-core configurations from either Intel or AMD - and, granted, I'm not so in-touch with computer hardware as I used to be - but my impression from the reviews I'd seen, etc. was that AMD was slightly slower per clock-cycle than Intel, and that the range of available clock speeds didn't go as high, either. The one advantage AMD had was a better memory architecture... Which, given the fact that this is a quad-core system, is very important if you don't want the problem to be dominated by a bunch of cache misses - but still my impression was that the Intel offerings were stronger at the current time...

Re:I've stuck with AMD

[Carbon016]

If an AMD box is $200 cheaper than a Intel box you are doing it wrong.

The E5200 is the same price as AMD's best offerings and smokes all of them. The motherboard is the same price, around $80.

Re:I've stuck with AMD

[TheThiefMaster]

Intel handled (I think they still do) all IPC (Inter-Processor Communication) through the FSB. Which is also the ram bus, and so runs at ram speed. They even did this for inter-core communication, completely screwing any attempts to scale a single intel system to lots of cores and have it still run well.
AMD have 3 independent buses, an inter-core comms for multi-core cpus (runs at cpu clock speed), hypertransport for inter-cpu comms and device comms (runs at multiple GHz independent of CPU speed), and an independent ram bus (runs at ram speed, obviously). This means that an AMD system gains real performance pretty linearly with the number of cores and cpus, and an Intel one didn't.

Intel countered by massively increasing their ram speed, countering the FSB bottleneck for smaller (2 or 4 cores) systems, and by making their cpus capable of more instructions per clock than AMD cpus (a real surprise when it happened), giving them great single-threaded performance. AMD couldn't match the performance of the most powerful Intel Core 2 cpus, so went for energy efficiency in a big way, and generally tried to undercut (instead of outperform) Intel at every turn. AMD's cache architecture was better too, with data not duplicated in all levels of the cache, so AMD cpus effectively had 10% more cache compared to Intel cpus. Intel countered by adding lots more cache to their cpus. AMD also went for forward and backward compatibility in a big way, a BIOS update (and sometimes not even that) is all that is needed to make the oldest socket 939 boards work with the newest AMD cpus. You lose out on a few features (e.g. faster HT and ram), but it makes upgrading an AMD machine much cheaper.

This leaves us with the situation where AMD cpus are great for highly-communicating parallel operations, and are great in clusters and datacenters due to having higher performance per watt (so they cost less to run and need less cooling). They also make for cheaper desktop systems both to build and to run, important if you're on a budget. Intel cpus are great for ram performance, and high-speed single-threaded ops, important if you are building a super-powerful gaming rig. Intel's pushing of their on-board graphics chipsets has also caused Intel cpus to end up in a lot of pre-built machines.

Though to be honest, You don't need a cpu costing more than £60 to play anything released recently at full speed, and AMD is incredibly competitive at those prices (e.g. my AMD X2 5600). The real expense is in graphics these days, though my 2-generation-old nVidia 8800 (rev 1) GTS 320MB hasn't struggled on anything I've bought recently, even on high settings...

Looks like the performance race might be slowing, unless someone comes up with a cheap, working holographic projector :)

Re:I've stuck with AMD

[Chandon Seldon]

from the reviews I'd seen, etc. was that AMD was slightly slower per clock-cycle than Intel, and that the range of available clock speeds didn't go as high, either.

That's all true. Intel's been beating AMD at that for a while now.

Thing is, that's never the question that you want to be asking when you buy a new computer. Who cares which company has the fastest chip at $1000. The important question is: If I spend $90 on a chip, what's the best I can get? What if I spend $150? Is that better than putting $60 somewhere else? How about $200?

In the $75 - $250 range (the range I personally care about), AMD and Intel are pretty much always trading blows. Here's a good chart for illustration: Crysis CPU Benchmark. Note how, for example, the Intel chip at $187 is slower than the AMD chip at $170.

Re:I've stuck with AMD

[petermgreen]

Who cares which company has the fastest chip at $1000.
only those who have money to burn and/or are buying more for penis length reasons than real value

In the $75 - $250 range (the range I personally care about), AMD and Intel are pretty much always trading blows. Here's a good chart for illustration: Crysis CPU Benchmark [tomshardware.com]. Note how, for example, the Intel chip at $187 is slower than the AMD chip at $170.
Note how the intel chip at $164 (rounding down to the nearest dollar like you appeared to do) beats every amd chip in the table in that benchmark.

To me that table seems to show one thing: intel THRASHING amd in the upper midrange. One overpriced intel chip (it appears the reason it's price is high is that none of the cheaper retailers stock it) does not change that.

Of course it will depend on your application, most of the AMD chips seem to be quad core so if you have applications that make better use of multiple cores things would look better for AMD.

Re:I've stuck with AMD

[Demanufacture]

Eventually Intel saw the error of their ways and they invented QPI: http://en.wikipedia.org/wiki/Intel_QuickPath_Interconnect

I'm not a staid AMD fanboy but you must admit that Intel realised that AMD had a superior interconnect technology and decided to take a leaf from their book (the more uncouth would call this "stealing" but AMD has borrowed their fair share of technology from Intel [x86, MMX, MMX2, SSE, etc] so you could say it's only fair)

Re:I've stuck with AMD

[Nalgas D. Lemur]

AMD also went for forward and backward compatibility in a big way, a BIOS update (and sometimes not even that) is all that is needed to make the oldest socket 939 boards work with the newest AMD cpus. You lose out on a few features (e.g. faster HT and ram), but it makes upgrading an AMD machine much cheaper.

Err, really? How's that? I know AM2(+) and AM3 are supposed to be relatively interchangeable, but Socket 939 doesn't have the same number of pins or layout as them, last time I checked. I would be totally thrilled if that were possible, since I still have a Socket 939 board, and my CPU is the only thing that's really not fast enough right now, but unless I'm missing something, I don't see how that would work. It seems like you'd just end up with some bent pins and a broken computer.

Re:I've stuck with AMD

[TheThiefMaster]

Whoops, I meant AM2 onwards :S

Re:I've stuck with AMD

[AbRASiON]

Ummm I don't know what you're talking about, the days of AMD being significantly cheaper are long gone and even if you do save money, the AMD performance levels are quite distinctively different to Intel now.

This is not the AMD Athlon XP thunderbird vs Pentium 4 days, where Intel was at least 2x more expensive for only 20 to 30% more speed, Intel is marginally more expensive for a lot more speed and a cooler overall system.

Furthermore back in the day, a good PC was 3000$ AUD, then 2000$ AUD then 1500$ etc - it's slowly decreased and decreased.
Buying Intel would be 3000$ and AMD 2200$ - now it's more like 150$ difference to get the Intel rig, it's really a no brainer.

I'm all for supporting the underdog (I have an ATI video card) but I mean sorry, AMD's products just aren't compelling and haven't been since the Athlon X2's when intel didn't have the core 2 duo out.

Re:Is GHz even important?

[NotBornYesterday]

GHz is one of several important factors in system performance. Speed and bandwidth of the bus is attaches to is another factor, as is the efficiency built into the CPU architecture itself (multithreading, cache size(s), etc.) to maximize the power derived from every available cycle.

Oblig car analogy: GHz = engine rpms, bus = transmission, cpu architecture = multivalve overhead cam cylinder heads with a supercharger strapped on top.

Re:Riight...

[poetmatt]

So wait, there's a magic processor I didn't know about that goes faster than two products that haven't been tested head to head yet, and it's intel's?

I have a bridge I'd like to sell you, too.

Re:Is GHz even important?

[LWATCDR]

We do have a lot about what is "underneath". We have seen the 45nm server parts and they are very fast per clock tick. The new Desktop CPU is based on the same design so we can work from that and the clock speed.

Re:Is GHz even important?

[Detritus]

That's assuming that everything is in cache.

Re:Riight...

[speedingant]

Its the ghz war all over again. Overclocking doesn't mean performance. Whoever modded me troll, whatever. CPUs are designed for the everyday user, to produce decent power whilst being cooled by air. I didn't RTFA, but for AMD to claim their CPUs are wicked fast * when cooled by LIQUID NITROGEN * is completely irrelevant. There are benchmarks comparing the Nahalem CPUs to AMDs offerings, and Intel laps them up.

I'll jump under your bridge when you jump under mine

Re:they'll beat intel on price, they always do

[Sj0]

All things considered, I think you'll have to be pretty cheap to care all that much about the power dissipation of your CPU. Even if it's 100W greater, you're looking at about 50 bucks a year at 6c/kWh if you're maxed out every day. That seems to me like the smallest cost involved.

Well, when you have free cooling...

[iris-n]

That's why I chose to work with superconducting magnets. Free liquid helium!

Although its pretty useless, I love having an overclocked desktop in my lab to show off to the simulation folks.

Now I can finally turn on Aero.

what a wank

[toby]

some people have got the 3.2 GHz offering up to 4.5 GHz on air

I bet that's super-reliable. Don't people have anything better to do?

Virtualisation should save AMD

[Jorophose]

Considering Intel only puts it on some chips, it's the only thing that makes me consider AMD right now.

That and the fact that 45nm Phenoms are coming to AM2+, so an AM2 board is not as "useless" as an LGA775... Sure, the LGA775 will get newer 45nm CPUs, and likely a few 32nm ones (or updated 45nm while the Nehalems are fixed up), but I feel you get better flexibility with AMD right now if you don't mind losing out a bit in gaming performance. Quads are also cheap when it's AMD. I'm just waiting for a 90W or 65W quad, and I'll buy one then.

What would be nice is more honest numbers, though. From what I hear the new CPUs (including X2-4850e) from AMD have messed up TDP; like Intel's old "meh this much I guess" measurements, nothing like the new Core 2 mesurements of "Danger! Danger! This shit's gonna blow!"...

But virtualisation takes the cake. I don't get a hit with AMD's CPUs, having Pacifica in all the chips.

Re:Virtualisation should save AMD

[aurispector]

I've been using AMD processors for years for two reasons. One is to keep the competition alive because it's a good thing. The second is because AMD processors are usually a better deal.

If you look at the benchmark charts over at Tom's Hardware, you can get a rough idea of how the processors rank in terms of performance. Then, when you go out an price them, the stuff that gives you the best bang for the buck is usually AMD. Not talking cutting edge or server class processors but the kind of thing you would actually buy to build a decent system without breaking the bank.

I have no grudge against Intel whatsoever - they're a great company with great products, but competition with AMD is one of the things that keeps them that way.

Re:Virtualisation should save AMD

[Jah-Wren Ryel]

But virtualisation takes the cake. I don't get a hit with AMD's CPUs, having Pacifica in all the chips.

Ain't no way I'm going to go dig up the links, but even with the latest virtualization support in Intel chips, the opterons are reportedly significantly faster changing from one VM to another. One hypothesis is that it is due to support for nested page tables on the AMD implementation, but I have yet to read anything more than empirical tests plus speculation.

Flops?

[tyggna]

Clockspeed is good and great and all, but that isn't a very good measure of how many calculations it can do. 6.0Ghz doesn't mean anything (especially if we're talking radio frequency) other than how quickly your processor's heart pumps. This is why the 3.0Ghz "barrier" has been in place for so long--it's because Intel internally switched from benchmarking by clock speed to benchmarking by floating point operations per second. Sure, it might be clocked at 6Ghz, but if I'm only getting a 6giga-flops outta that, I'd be better off buying an Intel Atom.

Re:Riight...

[Rockoon]

I guess you should actualy RTFA... then you wouldn't look like a moron.

I've stuck with discounts.

[Ostracus]

"First, I can put together an AMD box equivalent to an Intel for about $200 bucks cheaper. That money goes into my Video Card and I wind up with a better overall System."

Oh I don't know. Newegg has some nice deals on Intel/Mobo combo deals. Throw in any promo codes and rebates and you can get a good deal on an intel.

Clock speeds aren't doubling every 18 months now?

[VShael]

I bought a notebook about 5 years ago, that gave 2.4Ghz.
By the oft-misunderstood (I'm sure I'm using it incorrectly here) that would have meant that I could be buying a 9.6Ghz machine these days.

What am I not understanding here? (Bear in mind that CPU's and architecture are really outside my scope of knowledge/interest.)

Re:they'll beat intel on price, they always do

[WuphonsReach]

All things considered, I think you'll have to be pretty cheap to care all that much about the power dissipation of your CPU. Even if it's 100W greater, you're looking at about 50 bucks a year at 6c/kWh if you're maxed out every day. That seems to me like the smallest cost involved.

Or you simply care about noise. It's a lot easier to cool a 45W or 60W part then a 125W or 150W part.

And up here on Long Island, NY, we pay closer to $0.17 per kWh (Spring 2008 prices). So for a system that is used 2000 hours per year (basically business hours), you're paying $34 for every 100W. Or $149 per year per 100W if the system runs 24x7.

Re:Clock speeds aren't doubling every 18 months no

[mmkkbb]

You're misunderstanding Moore's Law. The number that doubles every two years is the number of transistors that can be placed on an IC inexpensively.

Re:Riight...

[Ant P.]

How else do you think they keep the planet balanced on that turtle!?