First Looks at Athlon 64 4000+ & FX-55

CrzyP writes "AnandTech.com has benchmarked the new "Athlon 64 4000+ and the FX-55" in various areas including business application performance, audio/video, gaming, and much more in this first look at AMD's newest 64bit chips. Just after AMD's announcement, AnandTech posted this article to help consumers choose between Intel and AMD."

Comments

wow

[Anonymous Coward]

I wish i had one

http://eric.blogdns.com/

Re:wow

[databyss]

I wish I could afford one :(

gmail from mbonig

[Anonymous Coward]

https://gmail.google.com/gmail/a-b0ab39f1a8-517235 8b19-6f45f145ca
https://gmail.google.com/gmail/a-b0ab39f1a8-13556c 0563-18969179a8
https://gmail.google.com/gmail/a-b0ab39f1a8-bc9b11 50e2-0bef3ba2a4
https://gmail.google.com/gmail/a-b0ab39f1a8-d6f302 cd83-e0644e7ef5
https://gmail.google.com/gmail/a-b0ab39f1a8-62e3c6 9d32-22621daaff
https://gmail.google.com/gmail/a-b0ab39f1a8-6c3f07 c84e-b9e70ce4cd

Re:gmail from mbonig

[Anonymous Coward]

Nice, even hosed Firefox...

Asshole.

Re:gmail from mbonig

[Anonymous Coward]

owned by GNAA

Oh Goody

[Anonymous Coward]

Now I get to run WinXP 64 bit's filthy code at 2.2GHZ! Hooray, worms propigate even faster!!!!

Re:Oh Goody

[Anonymous Coward]

Well, you're the one who's deciding to run teh 'Doze, so don't bitch to us.

uhm

[The Other White Boy]

AnandTech posted this article to help consumers choose between Intel and AMD

...what article?

Re:uhm

[Anonymous Coward]

There is nothing new there. No more info than in original press release

Re:uhm

[timts]

why they choose P4 3.4G EE? I saw some gaming machine company uses p4 3.8G OC to 4.3G(with water cool) as their machine's CPU.

IMHO, fair and square test should be based on CPU of same price tag(roughly).

Re:uhm

[Anonymous Coward]

Well, there's only 1 article linked to from the description...

You're right, this is confusing

Re:uhm

[Anonymous Coward]

Flamebait? Whoever posted the article forgot to add a link under "this".

Re:uhm

[The Other White Boy]

thank you. glad somebody figured it out. i kinda wanted to SEE this unlinked article.

Also here

[elid]

on Extremetech

Spread the love

[Hedon]

Why should Anand get all the attention?
Feel free to also check http://www.hardocp.com/article.html?art=Njc1

Re:Spread the love

[Anonymous Coward]

Probably because Anandtech submitted the article summary to Slashdot! What, you think these blurbs are actually written by people who do not have a financial interest in the website links or the products mentioned?

Re:Spread the love

[MortisUmbra]

Well, that and HardOCP sucks....gimme AT any day over HOCP.

Re:Spread the love

[teko_teko]

Because most of us don't like the [H].

Re:Spread the love

[berkut7]

Becuase, IMO, he's the least biased of them all.

Re:Spread the love

[Anonymous Coward]

Because HardOCP can't be trusted

my pc

[Alosja]

i am pretty sure it will do better than my old celeron(500 mhz), but who will need such an amount of computer power (except enterprises)? Of course new concepts will need more power(3d desktop?), but why are they deliviring this now?

Re:my pc

[Anonymous Coward]

...

....

..are you amish?

Re:my pc

[databyss]

Who is going to develop software that can't be run yet?

You need the power first.

Re:my pc

[BinLadenMyHero]

Well, id did that with Doom3.
They started writing it before the hardware capable of running it actually existed.

Re:my pc

[databyss]

#1) just because it wasn't in production doesn't mean it wasn't in the works. #2) Video game companies have to take risks like that to keep their games cutting edge, operating systems and other such software companies don't. There's no money in putting 200 programmers to work on an enterprise software suite that won't be useable for another 5 years.

Re:my pc

[Anonymous Coward]

You are a moron.

Re:my pc

[aposch]

Ever edited videos? Ever cut ads from a TV-record? This PC has 2.6 MHz, and boy, it's a slow box when doing such things ...

Re:my pc

[Anonymous Coward]

2.6MHz?? I'll bet it's slow.

Re:my pc

[Anonymous Coward]

Well, if you'd upgrade that Commodore to 64K RAM, you'd probably see an improvement.

I wonder

[robslimo]

With Intel having recently backed off on the effort to push clock rates ever higher, is there a plateau in sight for AMD? Will we not see anything between 5 to 10 GHz with today's techniques?

Maybe it'll take optical computing to spur the next clock push.

Re:I wonder

[stone2020]

AMD will probably max out at 3 ghz with this technology, but then they will introduce dual core which should be around 5 ghz.

Re:I wonder

[iggymanz]

clock speed isn't the only way to do more work in less time, and get better balance with the current bottlenecks getting to memory & i/o -- time for some new architectures, instead of the same old stuff with smaller transistors and higher clock speeds

Re:I wonder

[Kazrath]

Clock speed has been a really poor way to determine the speed of chips for awhile. The problem is that it is widley accepted that GHz=Speed of processor. As we have noticed with AMD running lower Ghz chips with faster throughput this is not the case anymore. Don't get me wrong Overclocking the chip and dumping more power into it will raise the Ghz/Mhz up and increase its speed. But that is akin to pushing the accelerator to the floor on your car. There is always going to be a max for the technology.

Re:I wonder

[Martin Blank]

Indeed. AMD's designs didn't scale as well speed-wise, so they had to get more creative to get better performance out of their chips. Improving their branch prediction, boosting FPU speeds, enlarging the cache, and improving the data bus were all methods that Intel would go back to from time to time, but usually with some reluctance.

Re:I wonder

[drinkypoo]

How about some new architectures with smaller transistors and higher clock speeds? :) I'd love to see, say, asynchronous logic, but I don't see it happening any time soon, at least in the main stream.

Re:I wonder

[swordboy]

With Intel having recently backed off on the effort to push clock rates ever higher, is there a plateau in sight for AMD?

There's some information to be realized:

AMD uses IBM's Silicon on Insulator (SOI) technology. This reduces power consumption by a very large degree. It is rumored that Intel tried to license the technology but, IBM and their fondness for cross-licensing, wanted too much (probably an x86 license). So Intel has been pushing out chips with standard silicon fabrication techniques at the expense of tremendous power consumption.

My guess is that Intel is coming up with a "massively parallel" architecture that can be applied to mainframes all the way down to handhelds simply by reducing the number of cores on a chip. The cores, will probably be very small and flexible. A mainframe might have a few thousand while a handheld might have a few dozen. They've certainly been hinting at a change in architecture for some time.

And then there was the "Windows Elements" that was supposed to come out with the P5. I'm not sure why that didn't get more press. I'm guessing that it is a version of Windows that will run in local storage on these processors (i.e. - the processor will have enough on-chip storage to hold "Windows Elements").

Re:I wonder

[ViolentGreen]

I think this is a lot of the reason AMD is starting to pull ahead of intel. Intel was able to just increase the clock speed to make their chips "faster" than AMDs. AMD had to look for other methods to increase performance and perhaps they learned a bit along the way.

Re:I wonder

[tanveer1979]

Here's are ready reckoner

130nm process, max 1Ghz
90nm process max 2.5Gz
65nm process max 5Ghz

Currently 65nm process is not stable in itself. Of course you can go for a 35nm process too but yields(working chips) will be Unless the yields are greater then 70% it does not make sense. Moreover as you push nm, electrical effects become more and more pronounced. For example electro-migration, crosstalk etc., We currently do not have the precision to go to such low nm. The next step seems to be optical, but along with it probably a 40nm process will appear.

Re:I wonder

[Anonymous Coward]

"130nm process, max 1Ghz"
???

130nm Northwood 3.2Ghz
90nm Prescott 3.8Ghz
65nm ???

Re:I wonder

[kc0dxw]

Microsoft would immediately take advantage of this by delegating the decision to blue-screen into hardware.

Re:I wonder

[Anonymous Coward]

Someone mod this dork funny. He went through the trouble to make his lame MS joke and will probably be hurt if everyone doesn't give him attention now.

Re:I wonder

[timeOday]

Are you claiming 2 x 2.5 ghz is worth 5 ghz? It isn't.

Re:I wonder

[stone2020]

I'm glad you already have that chip designed. I would call Intel or AMD because they will probably want to hire you. I understand that a dual core chip won't be twice as fast but I stated the limit of a single core chip would probably be 3 ghz. 3 X 2 = 6. So I subtracted 1 ghz but you also have to add speed increases for bigger caches (2 or 4MB) and 64-bit.

Re:I wonder

[Neuroelectronic]

Wow, thats some excelent math. Can I see your sources? oh you just made it up...

Re:I wonder

[PhaxMohdem]

A good theory however I would be led to believe that the mainstream software industry would be the main factor holding back the "massively parallel" cpus. Give the time it's taken for Microshaft to "work" on its x86-64 bit version of Windows I'd bet it is quite some time before your Joe Blow windows could support such a parrallel CPU(s). Let alone most 3rd part apps. Hell the vast majority of software available today still doesn't support multiple threads or hyperthreading, and that has been around for ages it seems. I guess necessity breeds innovation and only time will tell. Still a good theory.

Re:I wonder

[stone2020]

I did make it up based on things I've read on the web. Look at the original post. The guy was wondering what was the limit and that is my opinion. I've seen pictures of a dual core Athlon chip running at 2 ghz with 1 MB cache for each core. Imagine what they will have in two years. If you want sources, go to Google and type in AMD dual core.

Re:I wonder

[sumdumass]

and your point is what?

It was funy and somewhat true. Most of microsofts blue screens are attributed to hardware interactions. (at least the now anyways) microsofts blue screens typicaly only happen with somethign that accesses ring 0 memory areas and that is basicaly hardware or the kernel (as oposed to aplications in previous versions of windows).

Futher more it would be benificial if the blue screens were delegated directly to the hardware because the ability to trouble shoot it would be greatly increased. Also you should be able to create a diagnostic situation that could directly recal the errors making it easier when windows refuses to boot after the blue screen.

Re:I wonder

[tanveer1979]

Umm thats not 3.2Ghz. Just because a tiny bit of logic is running at 3.2 GHz does not mean chip is at 3.2 GHz. These are gimmincs semiconductor industry employs, and being an insider i know of it.

processors are great.

[ilovelinux]

just waiting for the right motherboard and I'm upgrading.

Does anyone know of a resonably priced mobo with PCI express (more than one slot), dual channel, and SATA? onboard lan vido and sound not reguired. prefer socket 939.

what's the best one out there performance wise?

Re:processors are great.

[bersl2]

Does anyone know of a resonably priced mobo with PCI express (more than one slot), dual channel, and SATA? onboard lan vido and sound not reguired. prefer socket 939.

You're gonna have to wait a bit longer...

Re:processors are great.

[Anonymous Coward]

yeah that's what I figured. how long until pci express is out en masse?

Re:processors are great.

[Frohboy]

Article number two on Anandtech right now is a first look at nVidia's nForce4 chipset, which (in the more expensive version) will support dual PCI Express video (to support the GeForce 6 series SLI capabilities). Essentially, there is a 16-channel PCI Express slot which can be split into two 8-channel slots.

They also make a passing reference to a chipset from VIA which will support PCI Express for the Athlon 64.

Toward the end of the article, they mention that nForce 4 boards from Asus and MSI should be out soonish.

Re:processors are great.

[Anonymous Coward]

let's add a few 64 bit PCI slots while we are at it.

Re:processors are great.

[erich.keane]

Actually, no, you really dont! TUL (AKA PowerColor) has released an AMD 64 939 board with PCI Express, based on the ATI chipset. Unfortunately, it doesnt seem to exist yet on their website, but i have some physical documentation that says otherwise:

A480A-VGF
Socket: 939
Chipset: ATI RS480+ULI M1573
Integrated VGA: ATI Radeon 9600
SYS bus: 2000MHZ
Memory: 4 DDR-400
PCI/PCIeSlot: 3PCI/2 PCIex1
PCIe16x: 1
Audio: Azalia 8 channel
PATA/SATA: 2/4
S/W RAID: 0,1,0+1
Ethernet: GB
IEEE1394: 2
USB2:8 ports
Form: ATX
Size: 305x245

There also exists a -VF, and a -V model:
-V: Same as -VGF, BUT Only 10/100 ethernet, and no 1394
-VF: Same as the -VGF, BUT only 10/100 ethernet

Re:processors are great.

[Anonymous Coward]

Does anyone know of a resonably priced mobo with PCI express (more than one slot), dual channel, and SATA? onboard lan vido and sound not reguired. prefer socket 939.

No, mostly owing to the fact that there's no such thing.

Re:processors are great.

[MortisUmbra]

Um, no, lets not....lets add PCI-E slots and get rid of PCI forever!

Re:processors are great.

[Royoken]

Well thats not entirely correct, nvidia has an announcement on there web page about the new nforce4 boards that will handle the new fx-55, SLI, dual channel, and the rest of the goodies

Impartial? hah.

[jkc120]

"Just after AMD's announcement, AnandTech posted this article to help consumers choose between Intel and AMD."

That's like saying "NBC has just finished a news story about Kerry and his stance on the issues, to help consumers choose between Kerry and Bush".

Or for you high school kids studying for the SATs:

AnandTech:AMD::NBC:Kerry

AnandTech and Tom's Hardware are so AMD-biased, I don't even bother reading them any more, especialy when it's related to the latest AMD or Intel chips.

Re:Impartial? hah.

[Anonymous Coward]

maybe it is just because AMD absolutely crushes Intel in every consumer arena.

cheaper, faster, no more heat problems

Re:Impartial? hah.

[Anonymous Coward]

bahaha! tom of tomshardware has always had his paycheque signed off by intel.

Re:Impartial? hah.

[Anonymous Coward]

You don't like biased sites and you still read slashdot?-D

Re:Impartial? hah.

[harrkev]

I certainly don't see that.

Well, in a way I do. They certainly are biased towards AMD, but only in the sense that AMD generally offers better value for the money, and is the only way to get 64 bits (right now). I bet if given the choice between a thousand dollars or a kick in the head, they would be biased toward the money ;)

All Intel has given us lately is a new extra-fragile socket, and PCI express (but good luck finding a PCI express vid card). In short, ho-hum. In a year or two, PCI express will be a good thing, but I will pass on it for right now.

Re:Impartial? hah.

[ergo98]

"AnandTech and Tom's Hardware are so AMD-biased"

Given the current political climate there seems to be a lot of ignorance over the difference between "bias" and "fact". Indeed often people confuse a lack of bias with a counter bias.

For instance, if indeed AMD is the superior consumer chip, perhaps offering measurably better performance/value, then it is entirely reasonable that a site would say such, and it isn't a "bias" to pronounce the AMD the superior choice of the current candidates. Similarly if George W. Bush dines on kittens for dinner, and someone reports it, that doesn't mean that they are biased against Bush.

Re:Impartial? hah.

[harrkev]

OK. I take back part my previous post. Intel HAS give the world hyper-threading. This is a good thing. Not as good as 64 bits, but significant.

Re:Impartial? hah.

[Kenja]

"and is the only way to get 64 bits (right now)."

Other then Sun, SGI, Alpha, Apple etc. Also, why do you need or even want 64 bit?

Re:Impartial? hah.

[leathered]

AnandTech and Tom's Hardware are so AMD-biased, I don't even bother reading them any more, especialy when it's related to the latest AMD or Intel chips.

Oh please, without any evidence or even anecdotes to back this up you're just making yourself sound like yet another disgruntled fanboy when reviews aren't going your way.

Almost all hardware sites agree that at the present, not only are AMD's chips providing the best performance, they also provide the most bang for the buck. That now even goes for media encoding, an area previously dominated by Intel. Who knows, no doubt the see-saw will swing back in Intel's favour in the future, but now it's clear that AMD will hold the lead throughout 2005.

Re:Impartial? hah.

[MindStalker]

But, if you ignore the fact that Kerry heats human fetuses, then you are biased! :)

Re:Impartial? hah.

[Jeff DeMaagd]

The thing is, Anand has been accused of an Intel bias, don't you remember the Xeon Nocona being pitted against a slower desktop AMD chip?

Re:Impartial? hah.

[jrod2027]

AnandTech and Tom's Hardware are so AMD-biased, I don't even bother reading them any more, especialy when it's related to the latest AMD or Intel chips.

This sounds so familiar to me... kind of like a MAC fan that won't even consider the competition as a viable option. Perhaps it is just you that is biased and not Anandtech and Tom's.

AMD has come out with some very solid CPUs lately, and frankly I think they deserve the recognition. When Intel comes out with its next flagship processor that topples AMD you can be sure that you'll hear about it on these sites.

Re:Impartial? hah.

[Anonymous Coward]

In the microwave?

Re:Impartial? hah.

[bill_kress]

In both cases I'd point out that it's possible that sometimes a more-or-less neutral group may look like it is supporting one side when, in reality, it is simply one side outperforming the other.

Is it really necessary to show both sides as equal when they aren't?

Re:Impartial? hah.

[MindStalker]

eats I meant.

Re:Impartial? hah.

[Jeff DeMaagd]

but now it's clear that AMD will hold the lead throughout 2005.

That is getting a bit ahead of things. 2005 is just as long a year as any other, and both AMD and Intel may have surprises in their sleeves, and either one might make a critical mis-step.

2005 is the year that Intel is planning to release dual core Pentium Ms, dual core desktop & workstation chips and dual core Itaniums. I haven't looked at the AMD roadmap lately, but AMD has dual core plans for 2005 too, to my memory it is just dual core Opterons, nothing for desktops and laptops.

Re:Impartial? hah.

[harrkev]

First, I am talking about x86 world here. Almost all games run on a Windows. Mapping software for my GPS runs on Windows. I am stuck with Windows (even if I dual-boot).

Second, a 32-bit processor is limited to 4GB of memory. I have 1/4 of that in my system right now. I can easily forsee that 4GB will be standard in about two to three years. To go further than that will REQUIRE 64 bits, unless you want to go to some goofy paging systems like we used to have to do back in the 80's.

Third, for some applications, 64 bits speeds things up a LOT. At the very least, the extra registers in 64-bit mode mean fewer loads and stores, which speeds things up.

Re:Impartial? hah.

[Kenja]

Intel has had memory controlers that allow up to 12GB of RAM on their Xeon systems for years. In addition, while yes a 64bit AMD system will support more then 4GB of ram, most of them have four DIMM slots. So unless your willing to pay an arm and a leg for 2GB memory sticks, your still stuck with 4GB.

In addition, 64 bit is slower then 32bit. The AMD systems run 32bit apps very fast, however this has nothing to do with them being 64bit. If everything else is equal, a 32bit system will out perform a 64bit system becuase it can address smaller memory blocks.

All that having been said, I ahve nothing against the AMD systems and in fact am planning on getting one. However I see nothing gained from it being 64bit rather then 32bit.

Re:Impartial? hah.

[Anonymous Coward]

Republicans have better sex lives, maybe he does eat pussy.

Re:Impartial? hah.

[leathered]

On the contrary I think I was a little conservative. The latest Intel roadmaps show that we will see the dual-core Xeon out in Q1'06 while AMD's dual core solution is slated for mid 2005. You must consider that Intel's dual core plans were only recently conceived after they realised that they've hit the MHz ceiling. AMD's 64 bit chips were designed with dual-core in mind from the start.

However, never underestimate Intel. They have resources, especially in R&D, that AMD could only dream about. It's still possible that AMD screw up and hand the initiative back to their much larger competitor but I still think that as far as performance goes, 2005 will belong to AMD.

Re:Impartial? hah.

[Anonymous Coward]

Competition?
What competition? /me sneers

Re:Impartial? hah.

[harrkev]

If everything else is equal, a 32bit system will out perform a 64bit system becuase it can address smaller memory blocks.

Huh??? This makes absolutely no sense. I know a bit about processor architecture. Going to a 64 bit address does mean that your instruction set must now allow for longer addresses, which MIGHT impact performance, but I can think of tricks to overcome this.

Intel has had memory controlers that allow up to 12GB of RAM on their Xeon systems for years. In addition, while yes a 64bit AMD system will support more then 4GB of ram, most of them have four DIMM slots. So unless your willing to pay an arm and a leg for 2GB memory sticks, your still stuck with 4GB.

True. But for Xeon, dump ALL of your existing software. Few apps, no games.
Your point about memory prices is well taken. In a sense, I agree. There is little to be gained RIGHT NOW from 64 bits. But RAM prices will fall, software will bloat, and in 5 years, 4GB will be an entry-level $399 system from eMachines.

Jumping up to 64 bits now adds very little to the cost of the system (performance-wise, compared to a similar system from Intel), and means that you are ready for 64-bit software when it arrives, instead of having to upgrade later.

Re:Impartial? hah.

[Kenja]

You claim to know about processor architecture but then claim that Xeon needs all new software. Xeon is x86. The current ones are based on the P4. Are you perhaps thinking of the short lived Intel 64bit CPUs?

Re:Impartial? hah.

[aicrules]

I've always felt that there is a similar slant on technology debate as with political debate. However, we (/.ers) are to processor choice as a senior political analyst is to a presedential choice. Throw all the charts and graphs at us that you want, we know how to interpret the real numbers and what they mean.

So yes, the uninformed public may start thinking, hmmm AMD sounds like a good choice if these guys say it performed well, but in all likelihood they will just go with whatever the sales rep tells them will fill their needs.

I guess the real question is....who is the NO SPIN for technology?

Re:Impartial? hah.

[harrkev]

Sorry. Brain fade. I was thinking Itanium.

Xeon is VERY expensive.

Re:Impartial? hah.

[Kenja]

"Sorry. Brain fade. "

No problem, happens to me all the time.

"Xeon is VERY expensive."

Wouldn't call them "VERY" expensive, but they do cost more. Starting price is 155$ for a 2Ghz with 533Mhz bus, the upper end is 459$ for a 3.2Ghz with 800Mhz buss, compare this to 179-585 for an AMD 64 3000 to 3800. Huh, seems the Xeon costs less, granted for most home uses its not going to be quite as fast. However most people dont need it. By the time 99% of the people would need or even want more then 4GB any system you buy today will be obsolete.

Re:Impartial? hah.

[jreilly]

AnandTech and Tom's Hardware are so AMD-biased, I don't even bother reading them any more, especialy when it's related to the latest AMD or Intel chips.

Man, Tom's Hardware is the site that did the comparison of intel and amd cpus playing quake3, while the heatsink was removed. The point being that the pentium3 froze, the athlons caught fire, and the pentium 4 kept going (though slower). People called them intel-biased then. More likely, they call it like it is. Amds did have serious heat problems, and now intels have serious performance problems.

Re:Impartial? hah.

[Anonymous Coward]

Too late. 0wned.

Re:Impartial? hah.

[crimson30]

Erm... DEC was first.

And besides... AMD has SMP. Same difference.

Re:Impartial? hah.

[Anonymous Coward]

Spoken like a true AMD fair weather fan script kiddie.

and tom's hardware

[he1icine]

another review on tom's hardware

http://www.tomshardware.com/cpu/20041019/index.h tm l

Re:and tom's hardware

[Sivar]

Surprisingly, that review seemed to be unbiased and made few if any rediculous claims based on miniscule performance differences. If they keep this up, I might start reading them again.

Hmm,

[hexag]

Yes, but will it Linux?

Like my boss said...

[Power Everywhere]

No one ever got fired for buying Intel. That's a shame since AMD seem to have better products and more innovative ideas.

Re:Like my boss said...

[Dot.Com.CEO]

That's IBM.

Your boss is an idiot.

Re:Like my boss said...

[RealAlaskan]

No one ever got fired for buying Intel.

Yep.

That's why ``... AMD seem to have better products and more innovative ideas.'' Since they're number two, they try harder. Once people have been saying ``No one ever got fired for buying AMD.'' for a while, expect them to stumble a few times.

Re:Like my boss said...

[Anonymous Coward]

I would challenge him to produce one example of someone getting fired for buying AMD...

Re:Like my boss said...

[Anonymous Coward]

really??

we lost a purchasing person at corperate because he bought intel.

we asked for some SGI workstations for a specific project. the nimrod decided he could save us $$$$thousands by buying Intel Based Dells instead.

He was fired.

Re:Like my boss said...

[Anonymous Coward]

moderation: -1: didn't get it
wish i wasnt too lethargic to explain your parent-post to you

Re:Like my boss said...

[Sweetshark]

No one ever got fired for buying Intel.
Really ? .... just wait a bit ...

Re:Like my boss said...

[ctr2sprt]

There's not a whole lot of significant innovation going on either at AMD or Intel, because that's not what people want. People want incremental improvements, so that's what they're getting. Even 64-bit CPUs have been around for almost a decade.

Most of the innovation going on in the CPU world right now is in the fab and design areas. We aren't getting innovative processors, we're getting innovative manufacturing techniques which allow us to do the same thing we've been doing for 30 years... just at a higher clock speed and with more transistors. And there, Intel is the clear "innovator," mainly because they actually have the money to build all the new-fangled equipment needed.

Re:Like my boss said...

[John Courtland]

IBM has been feeding AMD some great tech with in the last 2 years. I think enough to say that Intel, despite the monetary lead, has not out-innovated IBM + AMD.

Re:Like my boss said...

[upsidedown_duck]

That's the worst thing a purchasing person can do: second-guess the engineers' request. Nothing squashes morale worse than working with SGI/Sun/IBM/whatever for years only for some bean counter to declare that Windows on x86 is as good as UNIX for some task they don't understand. It is sad that so many vendors have jumped on the Windows bandwagon leaving some engineers with no choice but to put up with Windows' limitations because their preferred tool migrated because "Windows is the future". Just losing /bin/sh is enough to make Windows suck from an engineering point-of-view--installing Windows add-ons like cygwin only highlights what Windows lacks for performing useful work on large datasets and many files.

Anand does not know much about SC manufacturing

[bstadil]

CPU manufacturing is all about yields, if AMD can make more chips that work by increasing the die size by adding a larger cache instead of upping the clock speed, then that's the route AMD will take.

This is actually the last resort, as the cost of wafer real-estate versus speed increase is low. You rarely do this for raw speed rather for special needs like Servers and the like.

The increase in the speed for a workstation is probably one speed grade at a cost increase of 30% or so.

There is two good articles over on TheInquirer about Intels road map and why they have to go the Increase the cache route for 2005. Worth a read. Part One and Part Two

Re:Anand does not know much about SC manufacturing

[Anonymous Coward]

So you criticize someone else, and then support your position by linking to a worthless rumor mill? Do you have any links to REAL sources?

http://www.siliconstrategies.com

Re:Anand does not know much about SC manufacturing

[Martin Blank]

Wait, wait, wait, wait, wait, wait, wait, wait....

You criticize Anand, and then use The Inquirer as your data source?

That's kind of like criticizing a game commentary by Al Michaels, and then pointing to the wino on the street corner as your proof.

Re:Anand does not know much about SC manufacturing

[bstadil]

use The Inquirer as your data source?

FYI, spend 20 years in the industry only referencing TheInquirer.

Re:Anand does not know much about SC manufacturing

[Anonymous Coward]

Re sig:

Garbage in -- Gospel out

Garbage in -- "Garbage in -- Gospel out" out

More

[Anonymous Coward]

Another nForce 4 Review on Hardware Analysis.

Nforce4 also coming in a couple weeks..

[Anonymous Coward]

This is a major overhaul of the aging nforce3 chipset.. Check it out.

Expect a flurry of new advances by the end of the year.

I am ready to buy a new Linux system and am pulling hair out trying to make the best choice. Due to Linux compatibility issues (and mixed experiences with nforce2), I cannot really consider nforce4 so it will be Via for me. Though Nvidia will likely get the nod for graphics.

The 90nm chips are a mixed bag at the moment.

Re:Nforce4 also coming in a couple weeks..

[Anonymous Coward]

Ouch, say what? I couldn't imagine anyone not picking something other than VIA. VIA is the worst of all of them. VIA has more problems than I can count. Why do you think Newegg stopped carrying VIA mini-itx boards? Because they got so many returns that they were losing money.

VIA is teh sux0r.

Re:Nforce4 also coming in a couple weeks..

[freelunch]

VIA has more problems than I can count.

Here we are not talking about the MB manufacturer but rather the chipset at the heart of the MB.

For example, the MSI K8N Neo MB exists in a version with the NVIDIA nForce3 chipset and in a version with the Via K8TPro chipset. Just as ASUS uses both chipsets. Etc.

The gotcha in all of this is buying for Linux. All of these new boards and chipsets coming out is Just Great. Reviews that focus on Linux are a huge step forward. But if your primary focus is Linux, being in the early group of folks to give a new product the go can be a Real Bear..

Right now, we're about to see a whole new generation of faster MBs come out. At the moment, I still haven't figured out which MB/Athlon 64 combo is ideal and I've been wanting to place the order for a few days. As much as I want to wait for XXX to release YYY, I have to keep telling myself that my Linux requirement means it is the support date and not the release date that matters most.

Re:Nforce4 also coming in a couple weeks..

[lakeland]

Using the latest reverse-engineered nforce drivers (i.e. not the nvidia ones) I find nforce much more stable than Via ever was. YMMV, as always.

Power density

[lagartijo]

The power density at nm process (watts per square inch) has reached the nuclear reactors. See page 8. http://cnscenter.future.co.kr/resource/rsc-center/ presentation/intel/spring2003/S03USCQNS67_OS.pdf It's intel's but I assume it is the same for AMD.

Re:Power density

[Anonymous Coward]

It's intel's but I assume it is the same for AMD.

No, it isn't. Intel has many problems right now compared to AMD. Please don't try and foist Intel's situation on AMD. Especially outdated material.

Re:Power density

[lagartijo]

Hey, don't get me wrong. I love AMD, I have AMD and my brother owns AMD64. All PCs at my job are AMD. I did mention that because the review says: "If they had all of their cards lined up beforehand, the biggest unavoidable issue becomes power density, which you can't simply get around regardless of how mature your 90nm process is."

Re:Power density

[drinkypoo]

AMD processors just plain dissipate less power as heat than Intel ones do today, probably in part due to their lower frequencies. AMD has less heat problem (though not dramatically so) than intel at the same performance level.

Re:Power density

[Anonymous Coward]

It's probably better stated using the metric system:Wattage per millimeter. Remember the NASA fiasco? No the other one. Not that one either, the other moon landing debacle. That's the one.

More Links

[JaF893]

http://www.hothardware.com/viewarticle.cfm?article id=592&cid=1
http://techreport.com/reviews/2004q4/athlon64-fx55 /index.x?pg=1
http://www.bit-tech.net/review/364/
http://www.short-media.com/review.php?r=266
http://www.pcstats.com/articleview.cfm?articleID=1 666
http://www.tbreak.com/reviews/article.php?id=331
http://www.amdreview.com/reviews.php?rev=fx-55-400 0
http://www.techwarelabs.com/reviews/processors/amd 4000_fx55/
http://www.hardocp.com/article.html?art=Njc1
http://www.xbitlabs.com/articles/cpu/display/athlo n64-fx55.html
http://www.sudhian.com/showdocs.cfm?aid=614

forget the new CPU's

[Anonymous Coward]

Where's the damn motherboards that I've been wanting, namely ones supporting PCI-E with the AMD64? Stuff like the Abit AX8 with the Via K8T890 (ships when?), to say nothing of Nvidia's complete silence on the matter. I know there's an existing one for dual Opterons, but I don't feel like spending that extra $2000 for it just now...

Re:forget the new CPU's

[AliasTheRoot]

If you had read the article you would have noticed it prominently linked to a partner article on the Nvidia nForce 4 chipset, in fact they went as far as using an nForce 4 benchmark system:

Socket-939 Athlon 64 CPUs
2 x 512MB OCZ PC3200 EL Dual Channel DIMMs 2-2-2-10
NVIDIA nForce4 Reference Motherboard
ATI Radeon X800 XT PCI Express

Re:forget the new CPU's

[harrkev]

So?

These days, the chipsets provide features, NOT performance. As far as raw processing horsepower, the chipset is not even involved! For a benchmark which relies on a lot of disk I/O, I could see it making a difference (as chipsets provide the SATA/PATA interfaces). Check out the nForce 4 review. You will find that there is only about a 1% difference between different chipsets.

Re:forget the new CPU's

[AliasTheRoot]

The parent had asked specifically about nforce4 chipsets. The post pointed out that the article was a test of new Athlons on nforce4 chipsets.

amd bias?

[uniqueCondition]

if you're going to stand your ground with wintel and attack reviews from tomshardware & co. then i have to ask what you take issue with.

did you disagree with the test system?
the benchmarks used?

i've read tomshardware for years and have found them objective and informative. While their results disagree with your emotion you shouldn't make baseless remarks

Re:amd bias?

[Anita Coney]

I used to read Tomshardware until several years ago. It reviewed several 3d cards, including one from 3dfx. It tested them in OpenGL where the 3dfx card dominated. The reviewer stated that the 3dfx card did well in OpenGL because OpenGL was its "native" API.

Not only did the reviewer not know the difference between Glide and OpenGL, he didn't even know that 3dfx's advantage in OpenGL was due to the fact that the drivers didn't fully impliment OpenGL.

In other words, Tom's site is worthless.

Re:amd bias?

[Martin Blank]

I stopped reading Tom's after the nVidia bias issue. nVidia ads were plastered all over the site, and suddenly reviews that showed nVidia edging out its competitors in the benchmarks were showing that nVidia was 'clearly far ahead' and were 'the best cards available' even as Tom professed to have no bias. While eventually nVidia was head and shoulders above the rest (for a little while, at least), I just couldn't deal with Tom's blatant ignorance of his own bias at best, or outright lying at worst. I went looking around, and settled down with Anand, ArsTechnica, and HardOCP, with the occasional other thrown in. I still visit Sharky on occasion, who doesn't seem to have ever lost the principle, even if the site's not as informative as the others.

Re:amd bias?

[Anonymous Coward]

Toms' lost their credibility long ago when they came out blatantly pro-Intel. I got tired of reading the kiss-up language in their conclusions. I haven't been to their site in months, though I usually check it out from time to time to see the non-CPU stuff.

Anandtech is questionable as well.

Nowadays it's the Inq, the Tech Report, and Ace's for me.

Re:amd bias?

[ViolentGreen]

In other words, Tom's site is worthless.

I was fine until there. One bad review or a bit of wrong information does not make a site worthless.

Re:amd bias?

[Anita Coney]

If such an egregious and obvious error can get published, how could I ever trust the site to discover less obvious errors? Essentially, it'd be like reading the Weekly World News for real news: Pointless.

Re:amd bias?

[Brian_Warner]

Too easy. I no longer read Toms, the bias is simply too strong. Where, you ask? Take this example. In the same week, Tom abused ATI and nVidia (rightly so, IMO) for paper launches, then does a serious review of an Intel chipset that not only isn't available yet, but won't be for another 6 months!

A great quote, from the Xeon vs. Opteron battleground: "AMD can consider itself lucky, because due to the dual channel memory controller that is part of each processor, the dual Opteron has a nice advantage, despite having a clock speed that is 1.2 GHz slower."
Let me get this right, AMD is lucky that they designed the chip archetecture properly? This makes it seem like AMD are somehow 'cheating' Intel out of a win, I fail to understand how this is not bias.

Read the conclusions they come to, read into how they analyse the chips and you too will see the level of bias. Most of the benchmarks are fine, although they do seem to weight them towards the pro-intel ones, and I have no problem with the test systems used, but there is so much bullshit padding before and after the actual numbers that I no longer care what Tom and his minions say.

Re:amd bias?

[grondu]

If such an egregious and obvious error can get published, how could I ever trust the site to discover less obvious errors?

Then I guess you don't read newspapers or books, watch TV news, listen to news on the radio, or read news websites, do ya?

Re:amd bias?

[Anonymous Coward]

I stopped reading Tom's when I learned that Tom was a Muslim.

Re:amd bias?

[Anita Coney]

Don't you get it?! An internet site devoted to hardware and it didn't know the difference between Glide and OpenGL!

To analogize, I'd certainly stop reading any paper which confused the United State Supreme Court with the office of the President of the United States.

Re:amd bias?

[John Courtland]

At least newspapers publish retractions.

Re:amd bias?

[crimson30]

I stopped reading Tom's Hardware because of the constant anti-Intel bias. Read some of the conclusions of their Intel vs AMD shootouts. They generally have something negative to say about AMD.

I stopped going to Tom's Site when I read this:

"Summary: The P4 3.2 EE Wins 32 Times, The Athlon 64 FX-51 15 Times - An Uncertain 64-bit Future For AMD"

Uncertain Future for AMD? I'd like to look at benchmarks and draw my own conclusions, thank you very much. I don't need Tom's Hardware predicting the future of AMD's demise!

If it weren't for the biased conclusions and ad laden articles broken into several pages, Tom's Hardware would be great.

Re:amd bias?

[crimson30]

"constant anti-Intel bias."

Sigh... I obviously meant anti-AMD.

Duh

[igzat]

I think the choice is clear regardless of this article. Intel announced last week that they are giving up a 4Ghz Pentium CPU, and even the 3.8 Ghz model is very scarce. Where as AMD's Athlon 3800+ can be easily found, With the announcement of the 4000+ CPU, AMD has a clear lead over intel, and will until the Dual-CPU wars begin sometime next year. I think now is a good time to own AMD stock. Their marketshare is going to slowly increase over the next 12 months. I'm not taking sides here, just stating the obvious.

Re:Duh

[Anonymous Coward]

Except were system boards are concerned. AMD has been happy to sit back and let other people make boards for their chips. The problem is, many of the AMD boards out suck.

Intel on the other hand has a wide variety of quality system boards available. Good chipsets, good performance, and good features.

AMD is getting killed because of the current situation. Especially in the higher-end boards (like multi-proc SMP boards). I don't see it changing any time soon either. Sure you have the nVidia chipsets which are pretty good but the boards makers are turning out absolute crap.

Re:Duh

[cslarson]

I think you need to be careful anytime you recommend that people own a certain stock. A lot of the good news about AMD is already priced in to the stock. While I agree with you about their marketshare, and personally do own AMD stock, the best time to buy a stock does not necessarily equate to when all the good news is coming out.

Re:Duh

[Bryan Ischo]

That's some pretty tired anti-AMD FUD there. Can't you come up with any fresh FUD? FUD that's been out of date for 3+ years is just not worth even spouting.

Re:Duh

[Anonymous Coward]

Maybe this would be a good time for AMD to stop using the misleading 4000+ name. The chip doesn't run close to 4Ghz.

And if intel is increasing eficiency while keeping clock speeds under 4GHZ for the next few years, the 4000+ really won't say anything about performance.

However, on the marketing side, it'll still work as well as it aways has, in that people think the higher number is better.

How many people buying amd know that the performance rating isn't the clock speed?

-1 Lame Predictable Karma Whore Comment

[Anonymous Coward]

If anyone moderates this, or any similarly lame, comment up I will ensure that you never, ever get mod points again.

Apples to apples?

[Quixote]

I'm reading the review right now (I know, I should burn my /. membership card), and the first thing that jumped out at me were the difference in memory specs between the AMD setup and the P4 setup:

AMD: 2 x 512MB OCZ PC3200 EL Dual Channel DIMMs 2-2-2-10
Intel P4: 2 x 512MB Crucial DDR-II 533 Dual Channel DIMMs 3-3-3-12

Why not keep the rest of the components exactly the same, so we can have a _real_ comparison?

I'm no Intel fanboy (or an AMD fanboy, for that matter), but when you're doing such benchmarking, some attention to details would help.

Re:Apples to apples?

[Quixote]

Oops, my bad: I didn't notice the "DDR-II" in the specs.

Re:Apples to apples?

[Mdalek]

Also, note that the DDRII is clocked at 533 and the AMD is using DDR 400, they could not have made it any more equal - as there's no support for DDR II on AMD.

Re:Apples to apples?

[MindStalker]

I think the point was to find the fastest memory available for the system, this isn't just comparing processors its also comparing the possibly max speeds you can expect when using said processors.

Re:Apples to apples?

[TheSunborn]

Because that speed difference in ram access speed most likely is due to different design. The Amd chip includes a ram controller on the cpu which normally allowed it to access ram faster then the intel solution. My guess is that both ram chips are DDR3600.

But in general you should always test the system with the fastes ram that system support.

Re:Apples to apples?

[TheSunborn]

I did read the text again, and the reason they don't use the same ram, is that the intel board use 533 MHz ddr2 ram while the amd board use 400MHz ddr ram(PC3200).

Some "back on the envolope" calculations tells us that the ram on the intel board is fastes to return the first 128 bit of data, but that the Amd board is a bit faster, to read more then 128 bits in a row.

Martin

Re:Apples to apples?

[Anonymous Coward]

You use the best components available for each system for these comparisons. AMD and Intel use different memory technologies, so there's no other fair way to base the comparison.

Re:Apples to apples?

[JamesP]

Coz that's Intel pimping DDR-II. Yeah, it is slower, but watever, that's what Intel wants you to use.

Re:Apples to apples?

[NerveGas]

Because to use the 1066 bus, you need 533MHz memory. Although 533MHz DDR memory does exist, it's not at all standardized, sanctioned, or endorsed.

JEDEC has, of course, been far too slow to adopt new memory timings - as the looooong wait for DDR-400 to be adopted testifies.

steve

Re:Apples to apples?

[obeythefist]

That just wouldn't work, although it would be nice to compare that way, it would have biases towards certain processors for certain benchmarks.

A great deal of the processing power of a CPU nowadays comes from the strength of the memory controller. AMD's chief advantage is the integrated memory controller on the CPU. Intels main advantage is the speed of the memory controllers they use (allowing faster memory to be used than AMD at present).

So therefore, the performance, and selling points of these CPUs is based to some extent on the memory they can handle. What the benchmarkers do is use what they consider to be typical memory for the platforms they are testing with, to give a closer to real world performance benchmark.

Most importantly, however, due to the FSB speeds, it's simply not possible to test each CPU with the same speed memory in some cases. AMD can't do 533MHz, and Intel would have to underclock to do 400MHz. So there's just no choice.

Consumers?

[TimmyDee]

"Just after AMD's announcement, AnandTech posted this article to help consumers choose between Intel and AMD." So if by consumers, you mean people that read /. and hardware sites and not the general public, then yes?

Re:Consumers?

[timeOday]

Obviously they should switch over to growing banannas. Any fool knows that the average Joe will never read hardware review sites. But mmmm, everybody loves a good bananna.

Maybe some of you should read the article...

[adiposity]

The 4000+ isn't clocked any higher than the 3800+, it's just got a bigger cache. It's basically an FX-53; in fact, that's exactly what it is, sans the name. It would seem AMD is plateauing as well, but perhaps 90nm will get them out of the jam later on.

However, this is a wise move by AMD even if the rating isn't justified (hint: the benchmarks say it's not). Intel will never have a 4GHz CPU, and idiots who don't understand performance will see the 4000+ and want it because it breaks the 4000 barrier. It could backfire, but probably not, because even though 4000+ isn't justified, it's still faster than any of Intel's chips on 90% of applications.

-Dan

Re:Maybe some of you should read the article...

[C32]

Um, you can already buy "retail" 4ghz p4 systems from alienware http://www.alienware.com/ALX_pages/choose_alx.asp
Also, people regularly overclock 3.6-3.8 ghz p4's to 4.0-4.2..
I am certain that intel will have a 4ghz chip sooner rather than later, however non-oc speeds will probably hang around 4ghz for a while given current technologies.

Re:Maybe some of you should read the article...

[PhotoBoy]

Intel have publicly cancelled the 4Ghz P4, they're going the route of adding more cache to the chip (like the Extreme Edition) which will cost them a fortune because of the larger wafer size, this is because they cannot get the heat output of the P4 Prescott to a satisfactory level.

Overclocking a P4 Prescott requires water or freon cooling because of the heat. I don't think I've seen a system get to 4Ghz on air cooling and if someone did they'd probably need a larger than standard heatsink and noisy fan both of which Intel aren't going to ship with the CPU.

Re:Maybe some of you should read the article...

[mapmaker]

The 4000+ isn't clocked any higher than the 3800+, it's just got a bigger cache...It would seem AMD is plateauing as well

What about the FX-55 that was also reviewed in the article? Is that clocked higher than the 3800+?
Maybe you should read the whole article...

Re:Maybe some of you should read the article...

[W2k]

How is the 4000+ rating "not justified"? "4000+" does not mean "faster than a Pentium IV at 4 GHz" like some blockheads want to believe. It's just a name and it neither could nor should be used as an indicator of performance, other than relative to other AMD CPU's.

Let's do an analogy. We have BMW and Mercedes-Benz, both high-profile german car makers. BMW has a car called 320i, Mercedes has one called C320. They're pretty similar in size and price, but the Merc is significantly faster. Now think of Mercedes as AMD...

Re:Maybe some of you should read the article...

[adiposity]

It's not 5% faster than the 3800+, basically. It's 1-3% faster, which isn't surprising, since it's clocked the same but just extra cache. It's more like a 3850+, if we are comparing it to the older AMD64s.

In any case, it's bullshit to say the ratings aren't meant to compare to Pentiums. That's exactly what they are for. If Intel didn't exist, AMD wouldn't use rating numbers. They use them to make their products sound competetive to Intel's. Yes, I know 3000+ does not mean "exactly as fast as a 3GHz P4, etc., but in reality the model numbers are intended to be used by stupid consumers for choosing their chip.

The rating numbers mean nothing to performance geeks. They are there purely for marketing.

-Dan

Re:Maybe some of you should read the article...

[adiposity]

I did read it (and several others); yes, I know the FX-55 is clocked higher (2.6GHz). I was writing to those who praised AMD for reaching 4000+ as if the model number actually meant something, when in reality the clock speed didn't change a bit. In fact, it's just an old, rebadged FX.

In any case, that AMD has reached 2.6GHz means little, as the number of FX-55s shipped is likely to be low. Its high price guarantees that this won't matter much, but it does raise the question of whether AMD has yields to supply a lot of 2.6GHz chips. If not, AMD having the fastest chip doesn't mean much, as most of us will be stuck at 2-2.2GHz for budget reasons.

Yes, AMD did release a 2.6GHz chip, but will it be as hard to find as the Geforce 6800 Ultra? Its only available as a high-end, "luxury" chip at this point. I think they are having trouble at the high clock rates just like Intel, but it remains to be seen. Intel released a 3.6 months ago, but it's been nearly impossible to find until recently.

By the way, I'm running an AMD64 3000+ (754), which is basically a 3200+ with a little less cache. Here's hoping AMD can make some fast 90nm 939s, because I personally doubt dual-core will add much speed to my games.

-Dan

processors are great.... who cares?

[quintessencesluglord]

Yeah, but this is part of the problem. It's not about who has the fastest chip, but who has the best package (chipset, price, etc.) overall.

If I want PCIe and PCIx (or any combination in-between), SMP or single, 64- bit extensions or not; I have SEVERAL different options with Intel. With AMD, none, nada, zip.

And looking at the new chipsets that are supposed to be coming out for the Opteron/Athlon, that situation isn't going to change for a while (the new nForce doesn't even support PCIx).

Add to that the glut of Intel chips (and most likely the price reductions to follow); well the Nocona is looking pretty tasty right now (and about $200 less than an Opteron per pricewatch).

The lack of range for quality boards is hurting AMD. If you had no alliance to either chipmaker, would you choose some promise for next quarter or something that is functional and cheaper right now?

Re:processors are great.... who cares?

[Anonymous Coward]

> And looking at the new chipsets that are supposed to be coming out for the Opteron/Athlon, that situation isn't going to change for a while (the new nForce doesn't even support PCIx).

1. lots of opteron boards currently use PCI-X (look at IWill and Tyan offerings), they just use a 3rd party controller

2. take a gander at the upcoming IWill DK8ES (should be out RSN)-

2x AMD Socket 940 processor Opteron DP support
Chipset nVIDIA CKB04 AMD 8131
max 16GB Dual Channel DDR400/333
6 Channel audio
1x Gigabit Ethernet
1x PCIe 16x
1x PCIe 2x (supports SLI)
3x PCI-X 64bit/133MHz slots
1x 32bit/33MHz PCI slot
8x USB 2.0 connectors (4 external, 4 internal) IEEE-1394
6x SATA ports

Re:processors are great.... who cares?

[Anonymous Coward]

> And looking at the new chipsets that are supposed to be coming out for the Opteron/Athlon, that situation isn't going to change for a while (the new nForce doesn't even support PCIx).

um, lots of current opteron boards support PCI-X, they just use a 3rd party (AMD?) controller, so lack of support from the nForce4 is not an issue

you want PCIe and PCI-X? check out the upcoming IWill DK8ES (should be out RSN):

2x Opteron Support
16GB Dual Channel DDR400/333
6 Channel audio
1x Gigabit Ethernet
1x PCIe 16x
1x PCIe 2x (supports SLI)
3x PCI-X 64bit/133MHz slots
1x 32bit/33MHz PCI slot
8x USB 2.0
IEEE-1394
6x SATA

> well the Nocona is looking pretty tasty right now (and about $200 less than an Opteron per pricewatch).

considering an opteron 240 is $191 on pricewatch right now, i call shenanigans unless you have some source that pays you $8 for each nocona you take off their hands

Re:processors are great.... who cares?

[Anonymous Coward]

nForce4 is a consumer platform, so it doesn't contain server features.

And how is Nocona looking tasty? I don't even see the much-trumpeted 3.4GHz version on PriceWatch.

Flamebait indeed.

Re:processors are great.... who cares?

[quintessencesluglord]

Yeah, but that's my point. No pcie AND pcix. No options as far as single or SMP (Opteron only. I can get pcie and pcix on a P4. Nothing for an athlon). And thats not including DDR2, SSE2/3. Considering the implimentation of Iwill's last Opteron board, I'm not really looking forward to it (check out the thread on 2cpu). There are cheaper boards that can do the same thing on an Intel platform. 3GHz Xeon- $325 Opteron 250- $749 ~$200 average along all speeds The DK8ES hasn't been released yet. I repeat, nada.

Re:processors are great.... who cares?

[Anonymous Coward]

> No pcie AND pcix

did you read what i just wrote?

1x PCIe 16x
1x PCIe 2x (supports SLI)
3x PCI-X 64bit/133MHz slots

that's (2) PCIe and (3) PCI-X

> No options as far as single or SMP

what the hell does that mean?

opteron can run in single or dual or even 8-way
an athlon-64 can't run in dual, but neither can a P4, so what's you're point

> And thats not including DDR2

question: what has faster memory performance? a xeon with ddr2 or an opteron with ddr

opterons whip xeons in memory performance so bad it isn't even funny

> SSE2/3

if you run one of the few apps where this is absolutely critical, then fine, go with your precious xeons

(btw sse2/3 will be supported on opterons next year)

> The DK8ES hasn't been released yet

well you said earlier "that situation [no PCIe + PCI-X motherboards] isn't going to change for a while"

and i was correcting you, it is going to change, and soon

> 3GHz Xeon- $325 Opteron 250- $749

bad comparison, as shown in anandtech's review (http://anandtech.com/IT/showdoc.aspx?i=2205) an opteron 250 is more comparable to a 3.6 xeon

how much is a 3.6ghz xeon? well i'm not seeing anything under $900

Re:processors are great.... who cares?

[NerveGas]

If I want PCIe and PCIx (or any combination in-between), SMP or single, 64- bit extensions or not; I have SEVERAL different options with Intel

If you take PCI-E off of the list, you have just as many options with AMD. And because PCI-E is generally not as important in an SMP system ("I need twin GeForce 6600's in my 4-CPU system for network testing!").

Besides, if you added scalable memory bandwidth, then you really don't have any option with Intel.

steve

Re:processors are great.... who cares?

[quintessencesluglord]

True, but the concerns expressed by my customers isn't absolute performance. It is building something good enough with a clear path for future upgrades, and at several different price points/uses.

PCIe is just future-proofing (not specific to graphics). Several customers have expressed a preferance for AMD, but there are no boards which fit their requests. That leaves Intel.

Compiler Flags

[barcodez]

They test with a lot of binary only pieces of software. Wouldn't the compiler and the compiler flags that these binaries were created with make a hell of a lot of difference? - I don't know - I'm asking.

Re:Compiler Flags

[Anonymous Coward]

Of course, but in practice you only mention major optimizations that one of the processors might not support. For example you might mention optimization for hyperthreading or SSE2, or a total lack thereof.

Re:Compiler Flags

[twiddlingbits]

Yes,some difference could be attributed to the flags Based on Benchmarks I have done in the past, I'd say its maybe 2-3%. Nothing you would notice as the end user as the apps are pretty well optimized when released.

Never a 4 GHz CPU?

[CaptainSuperBoy]

I realize the Pentium 4 won't hit 4 GHz in its current incarnation, but is it really true that we'll never see a CPU above 4 GHz? Have we finally hit the limit?

Re:Never a 4 GHz CPU?

[adiposity]

We may, but not for a very long time, as they are dropping the speed down for dual core, probably. They ramped up to higher speeds than they should have, when they could have had the same performance with a lower clocked P3 derivative. Eventually they may get back up there, but the high clock rates are gone for a while, I think.

-Dan

Re:Never a 4 GHz CPU?

[drlake]

No big loss, really. What matters is output, not the spec description. OK, I'll clarify that, what matters to ME is output, not the label. I realize most consumers are too ignorant to realize that the numbers don't tell the story and don't even know where to go to get good analysis of the real differences. That means that it will be really interesting how they choose to market any new architecture, since they can't just go by speed.

Re:Never a 4 GHz CPU?

[FireBook]

it is a shame though, the p4, although doing so little (relatively) per clock due to such a ridiculous pipeline, would, nonetheless, have really been stratospheric if it had ever reached 5-6ghz+, pure output would have drowned everything else out. You cant blame Intel for taking the route they did, as the problems they incurred were not expected. The fact that they built waiting time into the pipeline to allow data to get across the chip in time before the next clock cycle really speaks volumes about how high a clock speed they expected the p4 to reach. (and iirc they patented the idea)

Re:Never a 4 GHz CPU?

[IncohereD]

The fact that they built waiting time into the pipeline to allow data to get across the chip in time before the next clock cycle really speaks volumes about how high a clock speed they expected the p4 to reach. (and iirc they patented the idea)

Exactly. I thought the original expectation for the architecture was something on the order of 10 GHz. Which would have been truly obscene.

this

[Anonymous Coward]

http://anandtech.com/cpuchipsets/showdoc.aspx?i=22 49

Is NBC biased, or just rational?

[Futurepower(R)]

Tom's hardware article: AMD's Athlon64 4000 and FX-55: Nails in the P4 EE's Coffin?. The article makes the point that, if you want to know which AMD part is the lower-powered 90nm part, you will have to read reviews.

NBC many not be biased; they may just be rational:

Government data compares Democrat and Republican economics.

Links to reviews of 3 movies and 35 books that say George W. Bush is the worst president ever: Unprecedented Corruption: A guide to conflict of interest in the U.S. government.

Duh Indeed

[goldspider]

Athlon 3800+ != 3.8GHz
Athlon 4000+ != 4GHz

I suggest you stop promoting something you don't understand. Educate yourself or you will sound like fanboy.

Re:Duh Indeed

[ViolentGreen]

The GP never said nor implied that

Re:Duh Indeed

[Anonymous Coward]

Yes he did. He stated the Intel 3.8Ghz is scarce while the AMD 3800 is available. He stated Intel dropped the 4Ghz chip while AMD will be releaing the 4000+.

He was directly comparing the two models and thereby implying that 3.8 and 3800 are the same clock speed.

Re:Duh Indeed

[SFBwian]

Or, you could alternatively read it as the comparable performance chip from AMD is being released, without Intel's usual similar offerings being widely available (or at all) at around the same time.

I think the GGP poster was giving the slashdot audience the benefit of the doubt with regards to their ability to disseminate an idea/opinion/fact from a statement.

Think. And while you're at it, assume others do too.

No Intel P4-4ghz

[BrookHarty]

I bet Intel rethinks about releasing a 4ghz P4 after this news.

Re:No Intel P4-4ghz

[Anonymous Coward]

Days ago Intel said they would NOT release a 4.0G P4.

Way to be up on the news smarty pants!

Re:No Intel P4-4ghz

[Anonymous Coward]

I said re-think, way to read a post!

Re:No Intel P4-4ghz

[SFBwian]

But... but... that would be flip-flopping!

Re:No Intel P4-4ghz

[NerveGas]

I doubt it. Their decision was undoubtedly economical - they just don't get enough yield to make it worth their while to sell them.

This competition from AMD will only serve to keep the prices down, so it makes gives Intel even less incentive to sell them. Had AMD not been around, they could have sold their 4GHz P4's for $2k each, and there would have been people who would buy them. But as it is, with AMD's "4000+" available, it puts an effective cap on how much Intel can sell them for.

Intel doesn't like to lose face. They don't like being one-upped. But in the past year or two, they've had to just suck it up and get used to it.

steve

Re:No Intel P4-4ghz

[BrookHarty]

I was thinking of it at another angle.

That the P4 4Ghz could easily be updated from the about to be released 3.8, and bridge the gap while Intel works on its new cpu's. I guess they could just let AMD have the show while they work on the next cpu's, dual cores etc.

P4 3.6 for 450 or a AMD 3800 for 600, Intel seems to be a little behind. Multi-core will really be the next step in power, prices havnt dropped enough, and the newer chipset motherboards are still not out. It's about to hit the next step when everything comes together, so soon....

Also, companies change their minds all the time, even after press releases. Hell, I'm sure Intel could release a 6ghz system if they wanted too. Some OC'ers already have aircooled that fast.

Re:No Intel P4-4ghz

[NerveGas]

Also, companies change their minds all the time, even after press releases. Hell, I'm sure Intel could release a 6ghz system if they wanted too. Some OC'ers already have aircooled that fast.

You think so? I'd love to see a link. I haven't seen anyone hit anywhere near that with air cooling. Someone might have once been able to get the figure "6 ghz" on a BIOS screen before the chip melted into slag, but that doesn't count.

In fact, I doubt you'll find a stable overclock of 6 GHz even with phase-cooling. Liquid nitrogen maybe, but nothing less. : )

steve

Dual-CPU wars

[corngrower]

From what an article on CNet stated, the dual cpu wars won't begin until 2006, as Intel has pushed back its expected ship date for their product. The article stated that AMD was expecting to start shipping in 3rd quarter of next year.

Good Link

[amitti]

Here's that link clickable and without the space:

http://www.tomshardware.com/cpu/20041019/index.htm l

no DDR2 support?

[Anonymous Coward]

Disappointing. Why doesn't nForce 4 support DDR2? It has every other conceivable bell and whistle...

Re:no DDR2 support?

[Anonymous Coward]

Chipsets no longer determine the memory technology on AMD platforms. AMD 64-bit processors have the memory controller on the processor.

Re:no DDR2 support?

[hawkbug]

This I realize - but my question is this: Say a new Socket 939 version of the Athlon 64 comes out supporting DDR2 - it wouldn't be possible for the motherboard to support it right? I wouldn't think so because of different voltages to memory, pin counts and mappings, etc. It would be great of DDR was pin compatible with DDR1 so it could fit in existing AMD based boards.

I've just been out looking...

[scribblej]

For some benchmarks that will help me decide whether I'm better off buying a low-end AMD 64 bit system or a high-end AMD 32-bit system. I've got about $300 to spend on a mobo/processor combination, and I see I can go either way at that price -- but as this system is only going to be used for playing video games (Half Life 2, specifically) I'm not sure there's any advantage to 64-bit. I see that it's got a better architecture even if you're only using it in 32-bit mode...

any hints?

Re:I've just been out looking...

[Confessed Geek]

I think the real question is will you be upgrading just your CPU next or doing another CPU + Motherboard upgrade?

If you are going to wait 2 or 3 years and go for a full MB+CPU then you can get the best bang for you buck on a high end Athlon or the "old" AMD64 chips. If you are going to upgrade your cpu again in a year then go for a 939 Motherboard and lowend 939 CPU like the 3500+ since you will be able to pop in a current FX chip a year and a half from now for the price of a current Athlon.

The other thing to take into consideration is will you need new RAM and Power Supply to go along with your new MB and CPU. Build that into your cost when you are comparing things.

I just recently spec'd my own systems and I went with the 939 AMD3500+ But I went with a whole new system since just about all my other hardware is 7+ years old ;) My thought on the matter was 2 fold in regards to 32bit vs 64bit.
1. I wanted to experiment with 64bit linux
2. I figured I would be getting a "free" upgrade when win64 comes out and 64bit games start becoming available.
2.5 Since I only do this once in a blue moon for my personal system, I wanted that warm glow of knowing I'll have a "little god" for a month or 2 ;)

Re:I've just been out looking...

[Anonymous Coward]

windows XP 64 is still months and months away.(if not another year)

Until then and you spend that extra $$$ for it you are wasting your money.

in about 6 months when XP64 finally does come out the system you spent $300.00 on will cost you $100.00

I say WAIT.

Re:I've just been out looking...

[Morkano]

I was in the same boat in the summer. I did a lot of looking around, and ended up going with a AMD64 3000+. It's pretty cheap, and the only difference between that at the AMD64 3200+ is that the 3000+ has half the L2 cache, which doesn't really effect the performance that much. And it's way faster than the XPs. It's really the on-die memory contronller in the 64 bit processors which makes them better than the XPs, and it seems to give them quite a considerable advantage.

Bottem line, I'd go 64 bit.

Re:I've just been out looking...

[mickes]

Well, if you read the Anandtech article you would have seen that on page 9 http://www.anandtech.com/cpuchipsets/showdoc.aspx? i=2249&p=9 is were the comprehensive game reviews begin. The second Benchmark on that page is the Half Life2/CS: Source Video Stress Test (VST) benchmark. The top of the line AthlonXP 3200+ (about the fastest socket A proc.) is 34% slower than the lowest performing Athlon64. That is all in 32 bit mode btw. On top of that, Valve has already anounced that they will be releasing a 64 Bit port of Half Life 2/Source some time after the release of HL2. Now Valve hasnt been the model of reliability but the possibility is there for a future boost to your performance. Finally, the Socket 939 platform has legs for the future, its gonna be around for a while. The Socket A platform is very close to the end of its life. The only real advantage that you get with Socket A is that mobo's can be found for cheap, however, if your like most gamers you probably dont want a generic cheap mobo so the cost here may be closer than you realize. Bottom line, pick up an Athlon64 3000+ or 3200+ cpu and a nice socket 939 board. Add a good mid-range graphics card like the 6600GT or the X700 Pro and about 512 of DDR RAM and you will have a very nice HL2 gaming rig. mickes

Re:I've just been out looking...

[scribblej]

Thanks to you, and to the others who have replied. I think now I have settled on a Socket 754 3200+.

First off, the cost difference between 754 and 939 is enormous! It looks like I'll double the price fo the proc and mobo just to make that jump.

Secondly, Morkano mentions that the 3200 has twice the L2 cache of the 3000 -- from looking around, it seems like there are two AMD 3200 chips; one is socket 754 and has 1mb of L2, which is sweet, and the other is socket 939 and has 512k.

I am already considering the cost of ram and a new power supply and a new video card on top of the mobo/proc combination, but I'm under the impression that the RAM won't need to be changed based on which mobo/proc I buy (they both will work fine with 400mhz PC 3200 DDR) and the video card -- well, I wanted a 6000-line nVidia, but my budget is probably going to constrain me to a 5700.

Thank again for the help!

Re:I've just been out looking...

[NerveGas]

Your best bang for you buck is a good NForce2 motherboard, a mobile AthlonXP 2500+, and a Zalman 7000a cooler. All together, that will cost barely more than $200, and it's highly unlikely that you'd get less than 2.3 GHz out of the chip - with 2.5GHz not unheard of. Not at all bad for a $200 setup! The Athlon64 chips are, indeed, faster - but the marginal increase in speed can cost you quite a bit.

Of course, if you want more future-proofing, wait until a good board with PCI-E is available, which will mean A64. Even though the extra bandwidth of PCI-E isn't a big factor now, the way in which video cards are able to interface with memory and the CPU are mroe intelligent than with AGP, and most all cards are going to move to it. Shoot, the 6600-series cards are enough to make you wish you had PCI-E.

steve

Re:I've just been out looking...

[NerveGas]

I think the real question is will you be upgrading just your CPU next or doing another CPU + Motherboard upgrade?

Upgrading just the CPU is rarely worth it. The performance advantage from better memory technologies and other interconnects (AGP vs. PCI-E) mean that upgrading the motherboard is usually a very good thing.

When you get down to it, if you run an Athlon64 in 32-bit mode, you're essentially running an Athlon XP with an embedded memory controller. Even at similar clock speeds, an A64 in 32-bit mode will still outperform an AthlonXP because of the lower memory latency, greater memory bandwidth, and increased cache.

steve

Re:I've just been out looking...

[Ahnteis]

Well, I just barely upgraded (from an AMD 2000+ -> 2700+) to hold me over until I could afford the REAL upgrade in 3-6 months. :)

Re:I've just been out looking...

[Anonymous Coward]

Don't you wish the diffrence were marginal

Re:I've just been out looking...

[NerveGas]

I don't have to wish. Just for fun, I just looked at a good number of benchmarks for an Athlon64 3500+, a $300 processer. I compared them to an Athlon 3200+ - although the overclocked mobile would be faster in both clock speed and memory bandwidth.

In very new, hardware-intensive games, the difference was as much as 30%. In office apps, MP3 encoding, Ogg encoding, and most other things, the difference was within about 10%.

So, if comparing a $300 CPU to a $85 CPU gains you only a 30% performance increase in the BEST case scenario, that's a marginal gain to me.

For an all-out gamer, if the extra $200 is worth a 30% improvement, that's fine - but I'd still argue that the extra $200 would be better put toward a better video card, and besides, the original author said that he was looking for something inexpensive.

I've got a mobile Athlon running at 2.3 GHz with a GeForce 6800GT, and I can tell you that there simply aren't any games that you can't play oh-so-silky-smooth with it. Even in Doom3 and Far Cry, with 4x AA and 8X AF, I can turn the resolution up pretty high and it's still amazing. I'd much rather have this setup than an Athlon64 with a $200 video card.

steve

Power Consumption - how much is waste heat?

[Bryan Ischo]

This very thorough article also includes a comparison of power usage of the various processors during idle and busy states. The numbers look HUGE - the 90 nm Athlon 64 3500+ does the best at 86 watts at idle, with the Intel P4 560 (3.6 Ghz) doing the worst at 124 watts. While under a workload, the range is 114 watts to 210 watts.

At first I couldn't believe my eyes - how can heat sinks keep up with these figures? But then I realized that only some of that wattage is being converted to waste heat - some of it is actually doing the useful work of the processor.

Just curious - does anyone have any idea what the likely waste heat dissipation, in watts, would be for these processors, given the total power consumption figures in the article?

Re:Power Consumption - how much is waste heat?

[BenjyD]

That figure *is* the power dissipation. There is no significant "useful work" that a processor can do.

Re:Power Consumption - how much is waste heat?

[Bryan Ischo]

I'm pretty sure that I read in a previous Anandtech article a couple of weeks ago that they are using a new means of measuring CPU power usage - they measure the total wattage consumed by the entire PC and somehow extract the CPU power usage from that. I could be wrong though, I'm not sure exactly how they are obtaining their CPU power usage figures.

I do not believe that the total wattage consumed by the processor would equal the total wattage produced by the processor as waste heat. The processor does more than just convert electricity to heat. It uses some of the energy provided to it to redirect electrons around its circuits. That energy is not lost as waste heat.

Maybe the amount of energy that a processor uses to do useful work versus the amount that it loses as waste heat is negligeable, I don't know. But I seriously doubt that these processors are dissipating 86 - 210 watts of waste heat. It just doesn't seem like it would be possible to build heat sinks and fans capable of dissipating that heat without serious impact on noise and form factor.

Re:Power Consumption - how much is waste heat?

[Chirs]

Nearly all the energy that goes into a CPU is dissipated as heat. Some of it is completely wasted, some of it performs calculations in the process. Consider that there are no moving parts other than electrons (which move quite slowly), the CPU doesn't make noise, and the output lines don't transmit appreciable power. You are putting energy in, so it has to come out, and the only output left is heat.

Re:Power Consumption - how much is waste heat?

[Bryan Ischo]

So basically, the amount of energy it takes to perform calculations is tiny? If processor A performs 2 billion arithmetic operations per second, and it is able to perform each operation just as efficiently as processor B which only performs 1 billion operations per second, the I would expect processor A to use twice as much energy performing its calculations as processor B.

But what you're saying is that the amount of energy being wasted as heat for both processors A and B is 99%, so the extra power used by processor A in its calculations won't be noticeable compared to processor B (assuming that the only extra power used by processor B is that used to perform calculations).

486s comsumed what, 10 - 20 watts? And they performed something like 1/100 or fewer as many arithmetic operations per second as today's processors? So they used 1/5 the power but performed 1/100 the amount of useful work. I guess that today's processors actually convert more of their input power to useful work (calculations) than processors of the past did.

Re:Power Consumption - how much is waste heat?

[nelsonal]

Almost all the power used in a processor ends up as heat because there arn't any moving parts (any that move flex and endup wasting energy as friction). There is a minute amount of power in a PC that goes to work (spinning drives and fans up and such) but usually this is considered negligable relative th the amount of power that generates heat. The assumption is that effectivly all power drawn is converted to heat through resistance (like a space heater). Calculations aren't work in the physical sense (application of force over time).

Re:Power Consumption - how much is waste heat?

[TheWizardOfCheese]

But then I realized that only some of that wattage is being converted to waste heat - some of it is actually doing the useful work of the processor.

All - 100% - of the energy entering a system must be either stored or dissipated. It makes no difference whether a system is mechanical or electrical. For instance, an elevator (if poorly balanced) could store energy when it is on the top floor, but it would lose it when it travels down again. The stored energy in one trip to the top is trivial compared to the energy input over many cycles.

Strictly, energy doesn't have to be dissipated as heat; it could be shed with higher energy photons. But that does not occur with CPUs. The bottom line is that every watt input to a CPU is shed as heat.

Re:Power Consumption - how much is waste heat?

[Anonymous Coward]

The processor does more than just convert electricity to heat. It uses some of the energy provided to it to redirect electrons around its circuits. That energy is not lost as waste heat.

Since processor does not doing any mechanical work (like lifting weights) and it does not transform any materials (like splitting water into H2 and O2), all it does from the energy standpoint is converting electric power to heat. So all 100% of the electric energy it gets ends up heating heating your room.

Re:Power Consumption - how much is waste heat?

[nelsonal]

The impressive part is that power generation occurs on such a tiny space, perhaps 1 cm^2. Also keep in mind that the cache is below average in heat generation so effectivly 1/3 of the processor space is generating probably 80% of that heat. That is a truly astounding heat density (several thousand watts/m^2).

Re:Power Consumption - how much is waste heat?

[Naffer]

Speak for yourself. I didn't have to buy an electric space heater or my room. I just play some ROME:TW for a few hours and I'm nice and toasty.

Re:Power Consumption - how much is waste heat?

[mczak]

you misinterpreted that power consumption graphs. This is total system power consumption, not only the cpu!
So if it says 100W, that is 100W measured at AC! Since psus are only 65-80% efficient, that means the system (without including psu loss) is only using 75W. If you keep in mind this includes hd, graphic card, mem, chipset,..., this doesn't leave that much for the cpu actually. Measuring system power also makes the differences in cpu power consumption look much smaller than it is in reality obviously.
And others have mentioned it already, ALL power is transformed to heat.

Re:Power Consumption - how much is waste heat?

[edwinolson]

It's virtually all waste heat.

The two basic energy-consuming activities are pushing charge around to charge and discharge capacitance, and leakage current. The latter is clearly waste; it performs no computation. The former can be conceptually be reduced to zero by reducing the parasitic capacitance of wires and gates.

If we consider more exotic ("abstract") computing systems, then it really becomes clear that we've got a long way to go before we start bumping up against any theoretical limits.

These papers might be helpful.

http://citeseer.ist.psu.edu/rd/27839242%2C24929% 2C 1%2C0.25%2CDownload/http://citeseer.ist.psu.edu/ca che/papers/cs/3548/http:zSzzSzwww.neci.nj.nec.comz SzhomepageszSzwdszSzfundphys.pdf/smith95fundamenta l.pdf

http://www.nature.com/cgi-taf/DynaPage.taf?file= /n ature/journal/v406/n6799/full/4061047a0_fs.html

Re:Power Consumption - how much is waste heat?

[NerveGas]

How can heat sinks keep up? If you've seen the size of heat sinks that come with these processers, you'll understand.

I built an P4/LGA system for a guy last week. The heat sink that came standard with the CPU really impressed me - it's the kind of heat sink you would have expected to see hardcore overclockers paying $60 for two years ago. Very large and well-designed!

Times used to be when heat sinks weighed one or two ounces, and came with 40mm fans. Then came the 60mm fans. Now, 80mm fans and two-pound aren't at all uncommon, with some models using 92mm fans, and some weighing three pounds or more. Copper is being used for more and more of the heat sink. Better heat conduction, more surface area, and more air. It's not rocket science. : )

Plus, on the new P4's, the chips are able to run at much higher temperatures than previous generations. The greater temperature differential between the chip and the heat sink, the faster you can get the heat to conduct.

steve

Re:Power Consumption - how much is waste heat?

[nmos]

Remember, what goes in must come out.

I do not believe that the total wattage consumed by the processor would equal the total wattage produced by the processor as waste heat.

I don't know what you're definition of "waste" is but it's true that virtually all of the energy used by the cpu ends up as heat.

The processor does more than just convert electricity to heat.

No, that really is about it.

It uses some of the energy provided to it to redirect electrons around its circuits.

Which generates heat.

That energy is not lost as waste heat.

If you ever make a processor that doesn't emit back out all of the energy that went in then get to the patent office in a hurry because we would no longer need seperate batteries for our laptops.

Re:Power Consumption - how much is waste heat?

[Anonymous Coward]

So basically, the amount of energy it takes to perform calculations is tiny?

Energy is conserved. The energy it takes to perform calculations isn't used up, it's converted to a different form. That form is almost always heat.

Take a ball, roll it across your carpet. You've put some energy into it. Eventually it comes to a stop, probably not where it started out. Assuming your carpet is level (and nothing got knocked over, etc.), all that energy became heat. It did some useful work along the way, but it still ended up as heat.

Some of the energy put into a processor might come out as electromagnetic radiation, and some might be used to redirect electrons to different pins. (redirection within the chip doesn't count, not until the energy leaves the processor)

We try to minimize electromagnetic radiation, because that causes interference with other devices (and other parts of the computer) and makes the FCC sad. So that usually gets converted to heat at some point, too.

Redirecting electrons (esp. at those voltage levels) takes very little energy, and most of that doesn't make it out of the computer. (until it's been converted to heat)

snicker

[DeathByDuke]

all these accusations of Tom's Hardware being AMD biased makes me laugh. Just be happy for once we dont have to moan about Intel biased websites anymore ;)

Re:snicker

[puto]

Oh ye of little knowledge of Anandtech.

His first ever review was of his hot dog cooker k6 chip. And the review appeared on his personal site that has since become the Behemoth that is Anandtech.

So he is partial to AMD, you could even say AMD was the piece of sand that turned into his personal pearl.

Puto

Re:snicker

[TCaM]

http://www.tomsphotos.com/router/

So now we know who did it.

Forget PCI-E

[Goosey]

Forget it, the AGP band has not been close to filled.. It will take years before we even approach a performance boost from PCI-E.

Re:Forget PCI-E

[spuzzzzzzz]

Forget it, the AGP band has not been close to filled

Absolutely. I remember a benchmark (I think it might have been on THG) which showed that AGP8x gets no benefit over AGP4x in gaming benchmarks. If the 4x bus isn't saturated yet, it will be a while before 8x is. I think this benchmark came out before doom3 though, so you might get saturation there with all the high-quality textures.

The main advantage of the PCI-E bus is the upstream speed. AGP8x is much faster downstream (ie. to the graphics card) than up (from the graphics card). PCI-E is the same speed in both directions. This would help video capture devices in high resolutions.

PCI-E is also intended as a replacement for PCI.

What distro can I load on it?

[Anonymous Coward]

But can I run Linux on it? The article doesn't even mention it. Why should I buy one of these if I can't even use it?

Wandering thoughts...

[Anonymous Coward]

AMD has their 'opteron' 8 way system.
8 x single core or 4 x dual core (mid next year), as I read it.

I'm gonna hold on to the old 533 celeron box and the 2 x 500 p3 box until the dual socket opteron mboards have 24 bit sound interated. Then 2 x 1 core and later 2 x dual core opterons (the fit the same board with firmware upgrade, according to AMD). I have no need for a single thread at 5GHz but aggregate 5GHz is appealing and this will get it, cheap. 2 x dual opteron is benching out at 26 or so for specfp.

This should be a great box for messing with DSP on desktop.

Curious. Other than possible cost savings, what is benefit of 5GHz single CPU with small cache over 4 x 1.2 GHz, each with lager cache and good switching/memory management?

Comment removed

[account_deleted]

Comment removed based on user account deletion

It's ALL "waste" heat.

[raygundan]

All of the energy going into the processor is going to come out as heat. It's similar to what would happen if you put a lightbulb in a box, and then measured how much heat was being produced outside.

Some of the power going to the lightbulb makes waste heat directly, and some of it makes light. But since it's all closed up in a box, all of the light ends up making heat, too.

So yes, some of the power going into the processor does useful work. But from the point of view outside the processor at the heatsink, even the useful work creates heat.

Re:It's ALL "waste" heat.

[NerveGas]

Well, even though it all gets converted into heat, it's not all wasted. Some of that electricity actually does some amount of processing, so you have received some benefit from the heat.

However, a very great deal of the electrical power is just leaked through the transistors in their "off" state, performing absolutely no useful work. I believe that for the 90nm chips, as much as 75% (!) of the electrical power is leaked, so you're really only getting anything useful out of 25% of the power you pump in.

steve

Re:32 bit Vs. 64 bit?

[RPI Geek]

*yawn* RTFA: At the end of the article there's a few pages dedicated to the bang-for-your-buck of each chip.

Re:32 bit Vs. 64 bit?

[NerveGas]

Those benchmarks were in 32-bit mode, so there's nothing to get alarmed about.

Note that simple 64-bitness buys you no performance increase, or even a slight decrease. Moving to the x86-64 64-bit instruction set, however, does provide an increase, but it's from the greater number of available registers.

steve

Sensible colours

[Anonymous Coward]

http://shit.slashdot.org/article.pl?sid=04/10/19/1 519249

Frequency != performance

[vlad_petric]

People like having a number that allows for quick comparisons between two products. Unfortunately, frequency is only one part of the equation.

performance = Frequency X Instructions per Cycle

Granted, if you keep the IPC the same, increasing frequency will help linearly. The problems with keeping IPC are: 1. complexity, which negatively impacts clock speed. 2. memory, which increases at a much slower pace (so if you miss out of the L2 a lot, increasing CPU frequency won't do much)

So, Intel said - people can't comprehend IPC, let's just ramp up the frequency. Their processors work ok for applications which are not memory intensive and don't have many branch mispredictions. AMD said - we can get better performance by improving the IPC (e.g. have a considerably better design for the memory system than Intel's). AMD processors would thus crush Intel processors working at the same frequency, but people don't really see that, because there's no aggregate performance number associated with the processor.

I personally like AMD's approach better, because higher frequency means higher power.

AnandTech posts this to help choose which is bette

[Petrus9373r]

Isn't AMD's current superiority to Intel somewhat obvious?

Re:AnandTech posts this to help choose which is be

[Anonymous Coward]

Intel and AMD both make products other than desktop processors, where Intel has some advantages. And let's not forget the whole SP2/AMD64 conflict thingy... Besides, most people dont't take statements like "AMD is better!!!!" at face value, they want to see pretty graphs and charts so they can try to make their own decisions. And that's why this is here....

II

Re:AnandTech posts this to help choose which is be

[Petrus9373r]

You are right, i stand corrected. I should not have been so generalized in my comment. I should have in cluded a phrase including something like "in the realm of processors". I thank you for pointing this mistake out to me