Intel's Pentium 4 3.4GHz Processors Reviewed

EconolineCrush writes "In one of the most gratuitous benchmarking indulgences I've seen, Tech Report has tested Intel's new Northwood and Prescott Pentium 4 3.4GHz processors against sixteen competitors ranging from the relatively old school Athlon XP to the opulent Pentium 4 Extreme Edition, with plenty of Athlon 64 action thrown in for good measure. Performance is tested in a wide range of applications, including gaming, rendering, image processing, media encoding, speech recognition, and scientific number crunching. Even if you're not interested in Intel's latest Pentium 4s, the review nicely shows where 18 of the fastest desktop chips from AMD and Intel stack up against each other."

Comments

The problem with all these new processors is

[Sir Haxalot]

once BTX comes out, they'll all be more or less redundant, unless the BTX manafacturers somehow make them compatiable, which I doubt is very likely.

Re:The problem with all these new processors is

[Anonymous Coward]

BTX is a bad idea. I only just got a new ATX case this thursday, which cost me around $100! I hope BTX fails.

Re:The problem with all these new processors is

[beswicks]

That is possibly one of the most bizzare comments to be marked interesting, just because a new form factor comes out, it doesn't mean that the processor companies will dump all the current chips.

True they may have a new package for some of the processors to fit a new slot or modified mb chipset, but that is nothing new, we don't just chuck out all the old work when something new comes along.

c.

Re:The problem with all these new processors is

[millwall]

they'll all be more or less redundant, unless the BTX manafacturers somehow make them compatiable, which I doubt is very likely.

And exactly would that be unlikely? Why wouldn't the BTX manufacturers want their motherboards to be compatible with the latest processors?

Re:The problem with all these new processors is

[binaryDigit]

once BTX comes out, they'll all be more or less redundant, unless the BTX manafacturers somehow make them compatiable, which I doubt is very likely.

How is that? BTX is a case "standard", what about a case is going to make a cpu "redundant"? How is Antec going to somehow make a cpu more "compatible"?

Re:The problem with all these new processors is

[freidog]

actually, we might see FMB 1.5 compatible chips (which newer S478 boards like the Abit IC7-Max3 should support), up to 3.8ghz by the end of the year.
Older 865 and 875 boards will only support chips for FMB 1.0 (limited to 3.2ghz).
Ace's Hardware has a nice layout.

Re:The problem with all these new processors is

[Slack3r78]

How in the world would that be the case? All BTX is is a new motherboard standard which rearranges the position of the components on the board to improve airflow and cooling. The CPU socket has been moved to the lower left corner of the board, with expansion cards in the top right, for example.

As far as CPUs are concerned, this is no different than having a Socket 370 board in ATX, mATX, and mini-ITX form factor. Could somebody mod the troll back down, please?

Re:The problem with all these new processors is

[Anonymous Coward]

Yeah, based on that logic, let's try another. "I just bought a new car this week. I sure hope no other new vehicles come out."

Re:The problem with all these new processors is

[Jexx Dragon]

I already hate BTX.

One, stuff is smaller. I prefer large cases with a lot of space for drives and cooling (my current project has twelve case fans).

Two, they are droping PS/2 ports, which I can't live without, since I perfer to leave the USB ports free (even though I rarely use them).

Three, everything is on the wrong side of the case, which means I'll not only have to get a new case to build one of these, but I'll have to get used of flipping it onto the wrong side to pop her open (Actully, the only way the left side opening is cool or good is one the Macs, where the side folds down to mount the mobo).

Four, What is this PCI Express?

Why not just expand the existing ATX standerd to include some new tech, wouldn't that be cheaper? Since the only set mounting location on the ATX standered seems to be the PCI and AGP slots and the nearly-universal dumb 12v connecter placement?

Meh, I'll still use ATX, Even if AGP video-cards are no longer supported (I do have a perfectly good command line, who needs fancy widgets and fonts?)

Re:The problem with all these new processors is

[jensend]

Prescott, PCI Express, and DDR2 are supposed to motivate people to move to BTX. Prescott is an absolute failure, while PCI Express and DDR2 don't show any real improvement over their predecessors. I wouldn't be surprised if Intel has quite a bit of trouble moving people to BTX; ATX will last a while yet.

Re:The problem with all these new processors is

[Zog The Undeniable]

Agreed. It's a wet dream if you make keyboards and mice though.

Re:The problem with all these new processors is

[HoneyBunchesOfGoats]

"we don't just chuck out all the old work when something new comes along."

But all the "old work" costs less! Maybe I'll be able to afford a good videocard in a few months. :)

Re:The problem with all these new processors is

[Joe U]

Like pretty much every other new technology standard, it's going to take a few revs before people get the full benefits of PCI Express and DDR2.

It will happen, the same way PCI made ISA video cards go away. It took a while, but it happened. Then came AGP, and PCI video went away.

Re:The problem with all these new processors is

[mikis]

This is a troll, right?

"I hate ATX. AT motherboards and cases are bigger, and who needs this fancy PCI and AGP when ISA and VESA are more than enough. Oh, and why replacing AT keyboard and serial mouse with this PS/2 crap?"

One, stuff is smaller. I prefer large cases with a lot of space for drives and cooling (my current project has twelve case fans).

You can fit small board in large case. But you can't put large board in small case. Most people DO NOT need big, ugly gray cases now that everything but CD and HDD is integrated on motherboard.

Two, they are droping PS/2 ports, which I can't live without, since I perfer to leave the USB ports free (even though I rarely use them).

Most of the new motherboards have at least 6-8 USB ports, and USB hubs are like 5$. So what is the problem?

but I'll have to get used of flipping it onto the wrong side to pop her open

Yeah, that will *really* be the problem.

Four, What is this PCI Express?

Replacement for both obsolete 32bit 33MHz PCI and AGP. Try fitting two top of the line (AGP) graphics cards on one motherboard.

Oh, and it is a good way to force you to replace whole machine when you just want to upgrade.

Re:The problem with all these new processors is

[Anonymous Coward]

PCI-X seems fine to me and is backwards compatible. I`d rather have compatibility _and_ increased performance rather than just increased performance.

Re:The problem with all these new processors is

[Anonymous Coward]

DDR2 will be gone before it gets those revisions (DDR3 is already on the market).

Re:The problem with all these new processors is

[jo42]

Did those 12 case fans come with a lifetime set of ear plugs?

Re:The problem with all these new processors is

[Dun Malg]

The point is irrelevant/off topic, he uses "redundant" to mean "obsolescent", and he split the first sentence of his post between the subject line and the coments. Off with his head!

Re:The problem with all these new processors is

[Bun]

PCI pushed out VESA and ISA within a processor generation. The same is likely for PCI-X. The benefits are just too great. PCI might linger on like ISA did, but I doubt it, since PCI-X is backwards compatible.

It took a while for SDRAM to push out EDO DRAM, too. It wasn't really until PC-100 came out that people actually started using it. The same will happend with DDR-II. When DDR tops out at 466 or 500, and DDR-II is at 566 or 600, the choice will be obvious.

Some people lack the perspective of history.

Re:The problem with all these new processors is

[jensend]

Benefits? What benefits? Except for GigE (who uses GigE anyway?) and high resolution video-in (do you have a need for real-time DV editing? didn't think so), almost nothing (on the desktop anyway, which is the primary PCI Express market) has any use for the increased bandwidth of PCI Express. PCI-X is an entirely different beast (PCI-X is almost the same as PCI but goes to 64bit 66mhz instead of regular PCI's 32bit 33mhz and is seen in servers and G5 Macs; PCI Express is the entirely different bus being pushed by Intel and is not backwards compatible by any stretch of the imagination).

Of course I remember the ISA-PCI transition. There were many more compelling reasons to switch to PCI, and the transition still took a very long time.

DDR2 won't last long enough for it to go through the technology revisions to become mainstream. It won't clock much higher than DDR either, largely due to power and heat issues. DDR3 is already on the market for use in things such as video cards and it runs quite a bit cooler at the same clocks than DDR2.

Re:The problem with all these new processors is

[mikis]

PCI-X seems fine to me and is backwards compatible. I`d rather have compatibility _and_ increased performance rather than just increased performance.

Please read this article, it is very informative: Introduction to PCI-Express: the AGP8X Replacement

To quote:
"The trouble is, while these technologies [66MHz PCI, PCI-X] have, or soon will find a permanent home in the server market, the complexities and extra costs they introduce to motherboard manufacturing mean that they will be virtually unknown at the desktop level. PCI-X, for example, requires a controller for every slot and that is just too expensive. "
...
"Approved as a standard on April 17 2002, PCI-Express is intended to be an evolutionary upgrade to the existing PCI bus. It will maintain complete hardware and software compatibility with all recent PCI devices."

Re:The problem with all these new processors is

[WuphonsReach]

Except for GigE (who uses GigE anyway?)

As soon as you start pushing DVD ISO images or AV files around like you used to push CD ISO and MP3 files around on your LAN, you'll quickly see the need for GigE. (Granted, it'll just move the bottleneck to a different point, but it's still a worthwhile upgrade.)

100Mbps LAN is capable of moving 17.5 to 26.0 gigabytes per hour across the network (at a roaring 5.0 to 7.5 Mb/sec). That means a DVD ISO image is going to take 10-15 minutes and a 50Gb tape backup will take 2-3 hours to move from A to B.

GigE, of course, can easily multiply that performance level by 5 or 7 (doubt you'll see 10x improvement, but you might). Hitting the disks on the server is almost as fast as hitting the local disks in my workstation. Pretty much a no-brainer upgrade now with 8-port workgroup switches for $150 on the street, managed 24-port switches around $1800, and GigE server NICs around $120.

The upgrade to PCI-X might be relatively smooth since it appears that a card can be designed to work in both style systems. (e.g. the website for 3com's GigE Server NIC indicates that it works with both PCI-X and 64-bit PCI.)

Re:The problem with all these new processors is

[Hoser McMoose]

One, stuff is smaller. I prefer large cases with a lot of space

The ATX specification calls for a maximum motherboard size of 305mm x 244mm. The BTX specification calls for a maximum motherboard size of 325.12mm x 266.70mm. I have no idea where you got the idea that BTX is smaller, because it's actually bigger. Maybe you're thinking of Micro-BTX or Pico-BTX? Those are both designed for small-form-factor designs, much like Micro-ATX and Flex-ATX today (the latter isn't very popular among consumer boards, but used in some OEM systems).

Of course, the size of the case does not necessarily have much to do with the size of the motherboard. As long as a case satisfies the requirements for the volumetric zones of BTX it can be any size the designer wants it to be. Full sized BTX cases will probably be pretty much teh same size as full size ATX cases.

Two, they are droping PS/2 ports,

Again, this is just flat out wrong. The BTX specification doesn't list ANYTHING to do with I/O ports except for the size of the rear panel I/O cutout. This is slightly shorter and longer than ATX, but not by a significant margin. It's still very possible to put PS/2 connectors on the back of a board, and I fully expect most motherboard vendors to do so. A few will opt to get rid of the PS/2 connectors in favor of more USB, but then again, Abit does that now with their "Max" line of ATX boards.

Three, everything is on the wrong side of the case

Err.. is that really that big of a problem?

Four, What is this PCI Express?

PCI-Express is a new point-to-point I/O protocol, allowing for a wide range of data rates. The main reason for PCI-Express is that it should eventually make everything cheaper by standardizing on a single bus instead of the multitude that we have now. Right now we've got PCI, PCI-X, AGP, AMR, ACR, CNR and CSA all being used on various PCs for a variety of uses. PCI-Express looks to improve on some of these and eventually to replace them all. Probably a good thing going forward, though not critical right off the bat.

Of course, PCI-Express has nothing to do with BTX. The first PCI-Express motherboards will all be ATX/Micro-ATX boards (maybe even a Mini-ITX board or two if we're lucky, the PCI-E 1x connector is very small making it a solution for tiny PCs). It's also quite possible to make a BTX board without any PCI-Express sockets, though that's a slightly pointless goal.

Why not just expand the existing ATX standerd to include some new tech

That's actually largely what BTX does. It's primarily just a redesign of the case to allow for more cooling for today's super-hot processors. ATX had the board sitting in the wrong place to do this, so a slight modification of ATX wasn't all that much of an option.

Re:The problem with all these new processors is

[cfuse]

True they may have a new package for some of the processors to fit a new slot or modified mb chipset, but that is nothing new, we don't just chuck out all the old work when something new comes along.

Shhh! Quiet, or I won't be able to pick up free computers from work in the next upgrade.

Speed

[Peden]

While all that processor speed is mighty good, who needs top-of-the-line equipment anymore? The new games all rely on the GFX card rather than the CPU. Any suggestions, other than the fact that Intel is keeping up to Moore's law?

Re:Speed

[jonjohnson]

If you end up building your own system, you're right. However, there are still plenty of low-end graphics cards that companies stick into computers just to save another 50 bucks in the manufacturing cycle. When this happens, you still have the "top-of-the-line" graphics chipset, but the board doesn't have its own processor. Without the onboard processor, the CPU does matter.

I remember a story in Wired a year or two ago that detailed how nvidia's CEO (or was it CTO.. it was a while ago) envisioned most of the workload for the computer on the graphics card, and the main CPU not needing to be very powerful. I wonder what he's thinking nowadays?

Re:Speed

[nacturation]

While all that processor speed is mighty good, who needs top-of-the-line equipment anymore? The new games all rely on the GFX card rather than the CPU. Any suggestions, other than the fact that Intel is keeping up to Moore's law?

Many non-game apps are CPU bound, and speed is always desired in these situations. Examples include rendering, video compression, SETI@Home, etc. Likely you don't need a faster processor, but it doesn't mean that the business world sees it the same way. Heck, maybe some day these processors will power your graphics card too!

Re:Speed

[Moonpie Madness]

I noticed that the Northwoods and older Anthlons were reasonably close in speed as well. And those medium range chips are getting pretty cheap! Plenty of great double proc. chipsets out there. If you want to blow the doors off of a prescott (that sounds do preppy) just get a double processor northwood for less. OR go with AMD, of course.

Re:Speed

[Wiz]

You should play Battlefield 1942 - it can use LOTS of time on AI. You also have other stuff like sound, network, inputs etc to deal with let alone something to keep the GFX card busy!

Re:Speed

[Peden]

I agree with you all, there is something to use the power for, no doubt. The thing I am missing is some KILLER app, like the time when Doom came it, that really revolutionized the industry. Now all it takes is pushing the graphics from no AA to 2xAA or from 1280x1024 to 1600x1200.

Re:Speed

[Anonymous Coward]

>Heck, maybe some day these processors will power your graphics card too!

no.

Re:Speed

[Kjella]

Many non-game apps are CPU bound, and speed is always desired in these situations. Examples include rendering, video compression, SETI@Home, etc. Likely you don't need a faster processor, but it doesn't mean that the business world sees it the same way. Heck, maybe some day these processors will power your graphics card too!

Not to mention, many of the games are CPU-bound because of the minimum specs - you can up the gfx from 640x480x16bit -> 1600x1200x32bit, but there's no setting the AI to "dumb -> average -> smart". I'm sure there's lots of interesting ideas in AI (groups, formations, tactics, responses to movement/sound, distractions etc.) or game world design (i.e. things happen to the world around you, not just what's being rendered on the screen) that'd love to have more power to throw at it.

Kjella

Re:Speed

[EulerX07]

This doesn't make sense. So you're saying that since some of the computers we can buy will come with a lousy graphic card, there's use for those cpus that costs 750+$?

How about buying the version of the computer with the 150$ cpu and switching the video card for a 150$ mid-end card from ATI or Nvidia? You'd wipe the floor with the 3.4EE computer with a lousy graphic card, and save 450$.

And also, how can you have both a "low-end graphic card" and a "top-of-the-line" graphic chipset? No offense, but the more I read your post the less insightful it gets.

Re:Speed

[Hecilwe]

While all that processor speed is mighty good, who needs top-of-the-line equipment anymore?

There are always research houses that need computations done as quickly as possible. If that's not convincing, keep in mind that as faster processors are developed, the market will drive down the prices of the ``lower end'' processors. It won't be too many years before you can pick up cheap Athlon64s for what an XP 2000+ costs.

Re:Speed

[Anonymous Coward]

"Heck, maybe some day these processors will power your graphics card too!"

No, some day you will use GPU for many things described above.

NVidia NV4x GPUs will support hardware MPEG1-4 encoding and decoding.

After a couple of years, you can also remove rendering from the list of CPU bound tasks.

Re:Speed

[walter_kovacs]

Exactly. Every day I have to run a script that generates randomized html pages - it's pretty straight forward, and I've optimized it as much as I can short of rewriting it in another language, but it is still pretty processor intensive and always maxes the CPU out to 99%. A faster processor is a god send, because it means that with a processor that's twice as fast, I have to spend half as many hours a day running this script and can use the cycles for something else. It's simple economics.

Re:Speed

[Moraelin]

Let me tell you about my day at work.

Just generating the RMI stubs for one small module with WebSphere's lobotomized Ant tasks takes anywhere between 1 and 3 minutes, depending on the module I'm compiling. On a Pentium 4 2.26 GHz. It's a purely command line process, so how would a GPU help?

Deploying anything on the Solaris test server takes ages. Deploying the GUI takes literally half an hour, punctuated by needing to click on the "next" button in that stupid wizard approximately once every 5 minutes.

Exporting the whole application from the test server takes 4 hours.

Could that server use a faster CPU? You bet.

And even in games, there's one factor which the marketroids at Nvidia and ATI conveniently ignore: the AI and physics.

All those screenshots may help sell a game, but people have started expecting more from the enemies than standing in place and waiting for you to shoot them. Or in flight sim you might have to not only run the AI for 20 planes dogfighting, but also a hefty physics engine to simulate their flight.

Then there are moddable games, and more interesting here are the scriptable ones. E.g., in "Morrowind", and even more so in "X2 - The Threat", a whole lot is programmed in an interpreted script language in the first place. The turrets' choice of whether to shoot an enemey or an incoming missile (and which enemy or missile), the freighters' choice which goods to buy from where, the destroyer captain's choice of which targets will get a broadside, those are all scripted.

Better yet, you can write your own scripts if you're a modder, or change the existing ones. Some people have done wonderful things with those scripts.

Run too many script heavy mods in either of the two games, and you may well discover why the CPU is still important.

Re:Speed

[jkcity]

games like moo3 and simcity 4 still utalise the cpu alot.

Re:Speed

[jonjohnson]

Firstly, the expensive processors always go down in price. Secondly, I'm talking about the average user, who doesn't know how to muck with their hardware. Sure, I can do it, you can do it, but do you really want every person under the sun trying to replace their graphics cards? Tech support nightmare... Thirdly, last time I went to the store, I had options as to whether I wanted to buy an GeForce FX 5900 without a processor on the board, or with a processor on the board. That distinction made the card go from roughly $130 up to $400 (in my particular Fry's store).

Re:Speed

[Katchina'404]

Thirdly, last time I went to the store, I had options as to whether I wanted to buy an GeForce FX 5900 without a processor on the board, or with a processor on the board. That distinction made the card go from roughly $130 up to $400 (in my particular Fry's store).

Damn, this is hot news, video cards without the graphics chip can now be bought from Fry's ?

Would you please direct me to a store where I can also buy said graphic chips by the unit, in a form factor that's easily pluggable into these "processor-less" video boards ? Thanks !

Either you're high on crack or veeery confused !

Re:Speed

[ionpro]

That's hardly true any more. Any modern game will allow you to set AI difficulty/AI CPU time, sometimes seperately. For instance, Battlefield (any variant) allows you to set "Overall bot difficulty" and "CPU time given to AI" (5~25%). I'm not entirely sure, but I believe Unreal Tournament 2004 has a similar setting.

Re:Speed

[Anonymous Coward]

and the fact that nobody here can even measure a decent difference between the low priced 2.8ghz model and the 3.4 overpriced model....

Call me when they hit 5Ghz.

only fools buy or even think of the highest speed processors...

Re:Speed

[gordyf]

Thirdly, last time I went to the store, I had options as to whether I wanted to buy an GeForce FX 5900 without a processor on the board, or with a processor on the board.

I think the salespeople were messing with you.

Re:Speed

[Anonymous Coward]

While all that processor speed is mighty good, who needs top-of-the-line equipment anymore? The new games all rely on the GFX card rather than the CPU. Any suggestions, other than the fact that Intel is keeping up to Moore's law?

I think this question comes up every time a CPU review article comes up. Anyway, the short answer is: the general public doesn't have much use for this power, at least not now. But there are always going to be people out there need more CPU power. That's part of the reason there are distributed computing projects... because they don't have enough computing power of their own. Also keep in mind that CPU industry isn't driven by games. It's a small (albeit important) market segment.

So what can you do with increased CPU power? Well, in the workstation/server world, you can either solve a more complicated problem, handle more concurrent jobs, or solve the same problem but get a better quality result (i.e. a simulation of a part where you can now factor in more surfaces and variables).

Re:Speed

[nelsonal]

It would be funny if there were a poorly sawn hole where the GPU is supposed to be. I think i have an emotion engine lying around here (from a dead PS2 can't recall what exactly died) if you want to try to try some reconstructive surgery.

Re:Speed

[shotfeel]

Examples include rendering, video compression,

And lets not forget the fact that some of us like to listen to our own background music while playing a game on the computer duing the time it takes for our favorited DVD authoring applications to encode the video.

Re:Speed

[Anonymous Coward]

Yeah, like I'm gonna buy a bleeding edge PC to run SETI@Home.

Re:Speed

[Hoser McMoose]

Nice thought except that there's no such thing as a dual processor "Northwood" (or "Prescott" for that matter). The P4 is simply not capable of running in a multi-processor setup, you need a Xeon for that. Now, one could make a pretty valid argument for a system powered by a pair of 2.4GHz Xeon processors, as those chips are rather reasonably priced these days, but the total cost of the chips + motherboard isn't likely to save you any money over these top-end chips. Better system? Perhaps (I'd certainly rather a pair of Xeons vs. one P4EE), but not cheaper.

The AthlonXP is a slightly different story. While it's not official capable of running in a dual processor setup, it can be hacked to work. Alternatively you could get a pair of slightly more expensive AthlonMP chips. Unfortunately dual-processor AthlonMP motherboards are rather dated now, so unless you're on a real budget a dual Opteron setup would probably be better.

Re:Speed

[obeythefist]

With a sufficient amount of experience with actual gaming, you'll find that your FPS are limited severely by the ability of your gaming rig to supply information to that graphics adapter. Overall gaming experience relies solidly on the efficient performance of each component that contributes to the game.

Generally the only real benefits gained from a better GPU are sustaining framerates at higher resolutions and graphics quality enhancements such as anisotropic filtering and antialiasing.

The reason for all this is mainly because there is a severe lack of optimisation in modern computer games. The CPUs and GPUs are so powerful that developers just throw code at them and hope it comes out quickly. The only optimisations that are done now are done for consoles because consoles don't have the power to pull it off otherwise.

Re:Speed

[Moonpie Madness]

well, color me a dumbass i guess you're right, i've been amd only for too long

Naming?

[Davak]

I thought Intel was killing their label of chips by speeds...

Davak

Re:Naming?

[nacturation]

I thought Intel was killing their label of chips by speeds...

Doesn't mean that the chips still don't have an internal clock speed. I guess this is the engineering benchmark rather than the marketing benchmark.

Re:Naming?

[mikis]

Enter Pentium 4 550.

Nice In-Place Ad

[Davak]

Thanks to Corsair for providing us with memory for our testing. If you're looking to tweak out your system to the max and maybe overclock it a little, Corsair's RAM is definitely worth considering.

Boy... I wonder how much memory Corsair donated for that wonderful little plug.

I can tolerate Coke planting their product in sit-coms... but I don't think I would appreciate my newscaster saying "Coke is so refreshing" in the middle of a news story.

Planting an obvious ad in the middle of "journalism" is just wrong.

Davak

Re:Nice In-Place Ad

[Indio_do_Xingu]

Yes I agree with you, but at least they said that Corsair provided them with RAM, giving everyone a hint on why their preference, not faking results as some sites usually do.

Re:Nice In-Place Ad

[Moonpie Madness]

Maybe it is cheesy and annoying, but commercial plugs sometimes make ends meet. I appreciated the content, and chuckled at the plug, but really, it didn't hurt anything. I assume that someone reading such an article would need more info than that to actually select corsair.

Re:Nice In-Place Ad

[binaryDigit]

Planting an obvious ad in the middle of "journalism" is just wrong.

What's so wrong about it? If they were testing RAM, then there would obviously be an issue. Or if the RAM were supplied by AMD or Intel. A manufacturer supplied a common piece of material for a comparison and the reviewers are acknowledging it, it doesn't hurt their credibility in one bit. Now if the review were sprinkled with "and such and such cpu performed so well on the memory benchmarks because of this wonderful Corsair RAM ....", then it would be questionable.

Plus your analogy of Coke being in the middle of a "news" story is off base. If it was a "news" story about how cola's rot the teeth, and Coca Cola donated all the drinks for an analysis of it's affects, then having the mention that CC supplied the drinks would be just fine (not that this scenario would ever occur of course).

Re:Nice In-Place Ad

[nacturation]

Planting an obvious ad in the middle of "journalism" is just wrong.

I don't see anything wrong about it. Imagine if you ran a tech review site and couldn't afford to equip all your various test machines with gigs of RAM each. Wouldn't you approach a company and ask if they could perhaps donate (or at least loan) you the equipment you needed? And, if they did such a thing, wouldn't it be nice to credit them for helping you out?

I fail to see how this is a "plant". It would be suspect if this were a review of sound cards and, right in the middle of the article, it said "Hey, your system needs more memory... purchase Corsair RAM today!" then that would be a plant. It would be no different than somebody comparing operating systems and thanking IBM/Dell/whoever for loaning you the equipment to do a side-by-side comparison with realtime parameter tweaking rather than having to tediously reformat a single machine every time you want to test a new config.

It's the lost art of the professional "thank you".

Re:Nice In-Place Ad

[JohnnyBigodes]

Shows up in pretty much every review they make. Corsair gives them the RAM, and it works fine. What's the big deal with plugging the brand in the review? It's not like it's some 2 minute TV-style ad jumping out at your face...

Re:Nice In-Place Ad

[Libraryman]

Boy... I wonder how much memory Corsair donated for that wonderful little plug.
I can tolerate Coke planting their product in sit-coms... but I don't think I would appreciate my newscaster saying "Coke is so refreshing" in the middle of a news story

This example is more akin to product placement than corrupt journalism. Likely Corsair gave them the ram for free, and maybe even threw a little ad revenue their way, in return they plugged the actual product that they actually used in the review.

Had they said "We used some ram we had lying around, but if you really want snappy performance try Corsair Ram" you would have the equivalent of newscasters plugging Coke during the evening news.

Besides, the evening news is no less entertainment TV than Friends, less entertaining, but not less entertainment.

Re:Nice In-Place Ad

[Anonymous Coward]

and this, my friends, is why all Linux users are teh ghey and are on teh spoke.

Re:Nice In-Place Ad

[222]

Corsair arguably makes the best ram in the world, and i have a hard time seeing why this plug is so shameless. If it were Spectek or something being plugged, I might be a little more cautious.

DISCLAIMER. I USE CORSAIR RAM IN THE PCS I BUILD, AND FIND IT TO BE HIGH QUALITY.

Re:Nice In-Place Ad

[shotfeel]

LOL. I was wondering if the line in question could be considered part of a "full disclosure" statement. In most cases it would be considered inappropriate to not state what hardware was used and where it came from.

Re:Nice In-Place Ad

[Anonymous Coward]

Corsair IS supposed to make the best RAM in the world, but I just went through 3 days of computer downtime due to an upgrade. I got some Corsair TwinX Dual Channel 3200 Platinum (2x512MB) for my ASUS P4P800 and P4 3.2GHz upgrade, and guess what, after some memory issues I check the MB manual, and for Corsair 3500 DDR is required. No thanks, I just got the OCZ 3200 Dual Channel kit instead, and saved myself well over $100 in the process. Fuck was I pissed. Idiots. Pissed at Corsair, and at the guy who sold me the stuff initially. My first call to the store to return the RAM was, oh yeah, there's a known issue with this MB-RAM combination. Duh! Why didn't you tell me this to begin with! Fucking idiots everywhere.

Re:Nice In-Place Ad

[lifespan]

The nature of the relationship with Corsair was disclosed so this is not cash-for-comment deception.

If you're looking to tweak out your system to the max

This hilariously lame marketing speak really hit my funny bone. I thought "to the max" went out with hypercolour tshirts, shoes with zippers on the side and terms like "rad", "gnarly" and "tubular". LOL

Missing 400Mhz....?

[inphinity]

Besides this test being ridiculously comprehensive, did anybody else notice the stat differences between the P$ 3.0 Ghz - 3.4 Ghz?

Or, more precisely, the lack of differences?
I wonder, is this just an inability of benchmark software to challenge a processor at such a high clock speed, or are these processors actually the same thing with shinier packaging?

Thoughts?

Re:Missing 400Mhz....?

[Anonymous Coward]

Thoughts?

Nope. Soz!

Re:Missing 400Mhz....?

[DigiShaman]

Some games (and 3D benchmarks) will be bottle-necked by the video card, leaving the CPU with spare CPU cycles to burn. Also, the benchmark may not require much general processing by the CPU, thus all the burden is pushed over to the video card.

Case in point. I was playing Warcraft3 on my P4 2.8 (with Radeon 9800 Pro). Though my framerate dropped down some at high resolution with 4x anti-aliasing, my CPU was only taxed at 15%. I noticed this after exit the game and looked at the task manager CPU usage stats. I was rather shocked.

Basically, if your a gamer, then your better of spending all your cash on a nice 3D card, RAM amount, then CPU ...in that order.

Re:Missing 400Mhz....?

[afidel]

Add to that the extra power usage 93W Thermal Design Power vs 103W and you have a real argument against using the faster CPU. My guess is that most of the tasks tested are bus bound more than CPU bound.

Re:Missing 400Mhz....?

[drinkypoo]

I disagree with the order in which you put your options slightly. I think memory is the single most important aspect. When windows kicks off some bullshit service task in the background memory is the #1 thing that will save you because if you are paged from here to nebraska (unless you're in nebraska, in which case you'll need to pick some point further away) when a new process loads it doesn't matter which task has priority, disk access will (in effect) go to the program paging you out. On many games, increasing their priority even breaks some timing loop or something - UT is one of them, but since it won't help anyway I'm not sure why I'm bringing it up.

Anyway everyone should have a gig of ram, but you can play any game out right now (though maybe not a couple games coming out soon) with a gf4mx. You can get a 5200 quite cheaply now as well, and that will run basically everything, though perhaps not at the highest resolution you might like. However, once you get to the point where you have an extra three hundred bucks to throw around, then by far the best way you can spend it (for improving gaming performance) is to buy a really slick video card.

As for the CPU, that is somewhat variable as well, if you play Unreal-engine games CPU is more important to you than to most people, they're typically CPU-bound. I'm not sure why that is, though certainly when Unreal came out the perception was that it had a more complex engine than its competitors, which let it do more things such as the wide open spaces we now take for granted. I mean UT2004 has enough room to fly jets in, as long as they're small and slow. Anyway it's almost certainly doing more math, so it's going to eat more CPU, which it does.

Re:Missing 400Mhz....?

[nelsonal]

I always though it was GPU, chipset, RAM, HD speed, then CPU for gamers.

Re:Missing 400Mhz....?

[SnappleMaster]

If you think windows commonly consumes a lot of memory with "bullshit services" you need to spend a few minutes looking around in task manager. And if you don't like the bullshit services, just disable them and stop whining.

RAM oftem helps gaming performance a lot but that's because a lot of games are memory hogs. E.g. you need 512 MB of RAM minimum to run a lot of MMORPG clients these days. These games don't need all that RAM just cuz they run on Windows - they need it because they allocate a hell of a lot of memory.

Once you have the minimum required memory, as dictate by the people who wrote the game, adding more memory is usually not going to impact performance a heck of a lot. So my ordering for where to spend your money on a gaming rig would be more like:

1) RAM, until the game can run optimally
2) Graphics card
3) More RAM, if it will help the game run better, YMMW
4) CPU

But that is assuming that your CPU is above the minimum specs required by the game to start with. If your CPU sucks ass then obviously you have a problem from the get-go.

Re:Missing 400Mhz....?

[SnappleMaster]

HD speed is way overrated by gamers. The games that I play do not hit the HD very much while the game is running. YMMV but if your game does not load a lot of stuff from the HD while the game is running then all a faster HD will do is speed up load time. That may be important to you but personally I value high framerates more than shaving a little time off loading.

Really Nice

[Mr. Certainly]

It's wonderful that we can create chips with so man transistors today. Only a decade or two ago, what we can build is truly amazing.

What are we building it for that is going to need that much speed at that price, and aren't there cheaper and more efficient alternatives at processing power than one uber-chip?

My 2 cents

Re:Really Nice

[Anonymous Coward]

What are we building it for that is going to need that much speed at that price
How about chip design. My lab just tried running Cadence(an integrated chip development environment) on a duel 2.4 Xeon and found it to be roughly twice the speed of a Sunfire V880, despite the Sun system having 8 times the memory and twice the number of processors. Others mileage may vary of course but for Design Rule Checking(DRC) and Layout vs Schematic(LVS) checks we saw 100% improvement(6 minutes vs 12 for one design).
aren't there cheaper and more efficient alternatives at processing power than one uber-chip
As far as parallel processing, this is a very difficult problem made harder by the fact that C/C++ has no implicit parallelism so the programmer must spawn seperate processes/threads to take advantage of multiple processors. Even then, there are algorithms that mathamatically cannot be done in parallel efficiently. Integrated circuit floorplanning is one of those problems and as such no tools are available that can use multiple processors.(We use our systems for other tasks as well that can take advantage of multi-processing)

Re:Really Nice

[Anonymous Coward]

Hi, I've got a complaint with your post.

I mean look, just because you're a human, and you're a member of the human race, and you have a brain, etc. that doesn't mean you should use 'we' when talking about what humans can do. Man alive!

Just leave it ok? You didn't have anything to do with this processor or most likely any other. You're just a person. You're not part of some collective pioneering force. You're most likey not pushing anything forward. You're not part of the gang. You don't work at Intel or AMD or anywhere, probably! It's THEM! It IS NOT YOU!

'we can create chips' - they can create.
'what we can build' - they can build.
'what we are building' - you ain't building shit!

Thanks for reading.

Re:Really Nice

[CuriHP]

6 - 12 minutes? You lucky bastard. My last DRC run was 9+ hours. Though I am running on a smaller box.

Re:Really Nice

[Glonoinha]

-My last DRC run was 9+ hours. Though I am running on a smaller box.

Get a faster box.
No, really. Get a faster box. Right now at Dell you can get the dual capable PowerEdge 1600sc with the Xeon 2.8GHz cpu (free upgrade from 2.4GHz), 1G ECC/Registered RAM (1G for the price of 512M), 18G 15Krpm SCSI drive, Gigahertz network adapter for $825. Buy two identical machines, port over the CPU, RAM, and drive from the second one and for under $1,700 you have a dual Xeon 2.8GHz box with 2G ECC/Reg memory and two 18G 15k RPM SCSI drives.

$1,700 is short money when you consider what it costs per day to run your shop - I'm not sure how much faster it would run (you didn't spec out your existing rig) but if you can get in two runs a day (4 hours per run down from 9) you have doubled the amount of work you can do each day.

Re:Really Nice

[Anonymous Coward]

That was only for small analog circuit, for quick testing purposes. One of our other circuits is a large array or digital circuits that we don't even attempt to simulate with HSpice or SmartSpice(1 million+ transistors). Its DRC takes 13-14 hours on our Sunfires. The Xeons we have lack enough memory (only 1.25 GB) for this design so it would lose to the Suns.
Intel has a very interesting paper on their verification process. I had it once and lost it. They use a simulation farm of about 1000 systems and do the chip piecewise. No full system analog simulation(as one would expect).

Re:Really Nice

[Anonymous Coward]

> aren't there cheaper and more efficient alternatives at processing power than one uber-chip?

Well, I think we have been seeing a general trend toward decentralization.

Graphics, video, and sound processing are now usually removed from the CPU and placed into their own specialized chips.

There is also an emerging trend to split processing up into parallel execution. We have seen the emergence of SMP (symmetric multiprocessing), where mainboards can support multiple CPUs, and the operating system can balance the load across them.

And within the CPU, we have seen increasing use of techniques like pipelining and multiple functional units that can truly execute instructions in parallel (often called "superscalar" execution).

All of these architectural changes were made to accomplish the goal you mentioned: a cheaper and more efficient alternative for achieving processing power.

Note also that sometimes we do want to retain the single uber-chip model, because of the physical limitations of transmitting signals from one place to another. At a clock speed of 3 GHz, the speed of light places an absolute maximum limit of 10 cm on the size of the CPU. At 30 GHz (achievable within the next 7 years), the CPU size limit will be down to 1 cm. At some point, uber-chips may become the only practical option for increasing speed.

Re:Really Nice

[Cynikal]

yup.. i still dream of the day where i'm so rich i can afford little wierd experiments.. like having someone build me a motherboard where theres a main 64bit cpu which delegates the processing to the IO cpu, and the graphics bus CPU, and the sound cpu, etc, etc... where ram is configued in a raid-0 aray for insane amounts of speed...

Re:Really Nice

[ChrisMaple]

High speed microcomputers for:

antenna design

aerodynamic and hydrodynamic simulation

organic chemistry (dna, proteins, viruses ...)

any complex system that needs detailed modeling over time, with analysis of the effects of different inputs.

Re:Really Nice

[drsmithy]

yup.. i still dream of the day where i'm so rich i can afford little wierd experiments.. like having someone build me a motherboard where theres a main 64bit cpu which delegates the processing to the IO cpu, and the graphics bus CPU, and the sound cpu, etc, etc... where ram is configued in a raid-0 aray for insane amounts of speed...

Any high end workstation machine does that *now*. No "weird experiments" needed.

Re:Really Nice

[CuriHP]

Unfortunately it's not my choice. It would get faster, but it would never be quick. This was over a million gates.

Re:Really Nice

[CuriHP]

The other problem would be the small amount of memory on that kind of setup. It wouldn't be the first time that 4G of RAM wasn't enough.

Re:Really Nice

[Cynikal]

you tell me where i can find a multi processor system where theres an AMD 64 main cpu, raid-0 ram configuration, p4E cpu taking care of all disk i/o with absolutely no over head on the main chip for ide drives, with another 3GHz processor taking care of all graphics processes, another one for sound, and another one for the rest of the I/O devices, and i'll buy it...

oh and a pci card or something with SDRam or DDR slots on it (for upgradability) to act as a ram drive (for use with the swap file, frequently accessed files, cache, and the sort)

Re:Really Nice

[drsmithy]

you tell me where i can find a multi processor system where theres an AMD 64 main cpu, [...]

Heaps of dual motherboards out there.

[...] raid-0 ram configuration [...]

Anything with dual (or more) channel RAM is providing the same thing.

[...] p4E cpu taking care of all disk i/o with absolutely no over head on the main chip for ide drives [...]

Any decent RAID controller. Not a P4E on board, of course, but it's hardly necessary - the i960s commonly used provide all the grunt necessary.

[...] with another 3GHz processor taking care of all graphics processes [...]

Any video card you can buy today will do its own processing on board. Heck, video acceleration hasn't been uncommon since the 80s.

[...] another one for sound, [...]

Any decent soundcard.

and another one for the rest of the I/O devices, and i'll buy it...

Again, decent server boards should have something like i2o. Also very common in high end non-intel unix workstations and servers.

Naturally, buying the best of all this stuff is going to cost you anywhere from 4x - 10x a "regular" computer, almost certainly with little, if any, real benefit - but the technology is certainly out there (and has been for a very, very long time - Macs had dual channel RAM, for example, back in the early 90s and video/sound acceleration has been around since at least the 80s).

Of course, you're never going to be able to remove *all* overhead from the main CPU - something has to tell all the other bits what they need to do - but the bulk of I/O processing, even in relatively low end machines, is not done by the main CPU (and hasn't been for some time).

However, any decent server/high-end workstation machine with hardware RAID, video and sound accelerators should provide with you with all the features you're after. You can buy all this stuff today. A bog standard off-the-shelf Dell Precision workstation will give you just about everything you want. A SunBlade 2500 will probably do it all. Heck, a ten year old SGI workstation would probably provide everything in your "wish list".

In any event, it'd mostly be a wank. I/O generally isn't slow because of the processing overheads, it's because the I/O devices themselves are slow. Your hard disk isn't slow because of the CPU overhead involved in accessing it, it's slow because the platters and heads can only move so fast. Similarly, your video card (generally) isn't limited by the CPU overhead in telling it what to draw, it's limited by the GPU.

oh and a pci card or something with SDRam or DDR slots on it (for upgradability) to act as a ram drive (for use with the swap file, frequently accessed files, cache, and the sort)

Seems a bit silly to do this when you could just put the RAM into the main machine, but you can buy such things now if you want to. For example. I've also heard of products which sit in a drive bay and attache to a SCSI or IDE controller to provide similar functionality.

Re:Really Nice

[Cynikal]

ok, copy a large file from one ide drive to another, or even one partition to another, while watching your cpu usage, and when it spikes to 98% (give or take) then tell me that again.. i wasnt commenting on the io being slow, but on the huge bog in the system during moderate disk access. and you can correct me if im wrong, but i was lead to believe that even a raid controler wont reduce the cpu overhead by much. of course, never having owned one, all i can do is repeat what i've heard and talk out of my ass.

i know some of the things i mention are sort of available now, but im not talking about having a work station, im speaking of an actual i386 system.. im talking about more power. im talking about upgradability, being able to add faster CPU's and more ram to any part of your system where it may be lacking. graphics not performing well enough? get a faster cpu for it and throw another 128mb ram in it without having to buy a whole new videocard.

now, gfx accelleration? i really mean i would love to see a gfx card running a 3ghz GPU. what they have out now is geat, but im a power user, so that means my phylosophy in life is "if its on the market, its not fast enough anymore" i want more.

if you dont know what i want the "silly" ram drive for, then you obviously havent used windows.. you'll find more performance comming out of a windows box if it has 512 megs ram and its swap file hosted on a ram drive than it will with 1.5 gigs of ram and a regular disk swap, or even dissabled swap. windows is very sloppy when it comes to managing its ram, and depends on its swap file, and thats when your bottleneck happens.. you can dissable the swap file, but then you'll need to either reboot every few days to clear the ram, or not use your computer for anything. the ram dive gives you the speed you want of ram, but tricks windows into using it responsibly. Plus i might want to use my entire (hypothetical) 2 gigs of ram and still have a ram drive to boot. and maybe im a little crazy, but i'd love to try booting my os directly from a battery protected ram drive.

yeah yeah dump on me for being a windows user, but for the whole combination of things i do, i find im more productive if im working rather than trying to figure how to get work done.

but thanks a mil for that link, i never heard of such a product actually being sold, i think i'll add it to my irresponsibility-binge list. but the idea of having the actual slots was more of a recycling idea to re-use old ram.. i already have sticks of pc2700 collecting dust, and i just feel like im throwing away money if i sell it for the going rate (since i paid way more than that for it) i'd feel better if i was using it for something.

ok and one last note here, being a musician who does alot of multi track recording on my pc, dont tell me that my sound card accelleration is "good enough", just try to lay down another track on a multitrack project playing back 12+ stereo 44.1khz wav's at the same time. for those who dont use their pc for recording, imagine a good ol game of unreal 2003, with all the gfx maxed, and 62 players jump into your FOV. (maybe im exagerating, but not by much)

and the final question you must be asking of why the hell would i want all that if i was rich, i could drop the ide drives and go to scsi, i could afford the newest video card the minute the prototype is built, etc, etc.. but my ideas arent just for me, its for everyone like me, who cant afford the best all the time, and harbors a DEEP resentment to chip manufactures, mother board designers, and ther rest of the companies who change 3 pins on their design and so my next upgrade will require a whole new motherboard, ram, cpu, and psu, etc, etc.. know what i mean? innovation in the areas i was thinking about would benifit everyone. (except monopolistic companies who extort needless hardware sales through jumping from slot to socket to socket with 2 more pins, to cresent shape cpu's etcetera)

Re:Really Nice

[Cynikal]

my GOD i never shut up, do i?

sorry

Re:Really Nice

[drsmithy]

ok, copy a large file from one ide drive to another, or even one partition to another, while watching your cpu usage, and when it spikes to 98% (give or take) then tell me that again..

You need to turn DMA on. This shouldn't happen, even with IDE controllers from the 486 era. If it is, something in your system is broken and/or incompatible with the rest of it.

and you can correct me if im wrong, but i was lead to believe that even a raid controler wont reduce the cpu overhead by much. of course, never having owned one, all i can do is repeat what i've heard and talk out of my ass.

No hard disk access in this day and age should have particularly high CPU overhead, regardless of interface. Certain IDE configurations (more than one drive per channel) will certainly involved a great deal of I/O overhead on the controller side - which may appear (depending on the tool being used) to be high CPU usage, but it isn't really. SCSI is marginally better in terms of CPU usage, but even IDE isn't _bad_, if configured correctly.

i know some of the things i mention are sort of available now, but im not talking about having a work station, im speaking of an actual i386 system..

Motherboards with most of these features cost under US$100. Even a relatively low end machine these days can be configured with dual channel memory for little extra cost.

im talking about upgradability, being able to add faster CPU's and more ram to any part of your system where it may be lacking. graphics not performing well enough? get a faster cpu for it and throw another 128mb ram in it without having to buy a whole new videocard.

There's no point. It would cost more to buy an "upgradeable" video card (or whatever) and incrementally upgrade it than it would to replace it every couple of years. Heck, technology in motherboards generally moves so fast that buying a motherboard with plans to do anything more than a trivial upgrade is a pointless exercise - and video card technology moves much faster than that.

now, gfx accelleration? i really mean i would love to see a gfx card running a 3ghz GPU. what they have out now is geat, but im a power user, so that means my phylosophy in life is "if its on the market, its not fast enough anymore" i want more.

I could nearly guarantee you any GPU on the market is faster (for graphics) than the fastest P4 available. Clock speed is not everything.

if you dont know what i want the "silly" ram drive for, then you obviously havent used windows.. you'll find more performance comming out of a windows box if it has 512 megs ram and its swap file hosted on a ram drive than it will with 1.5 gigs of ram and a regular disk swap, or even dissabled swap.

I use Windows a lot and I sincerely doubt that is true. I'd be interested if you have any benchmarks, however.

windows is very sloppy when it comes to managing its ram, and depends on its swap file, and thats when your bottleneck happens.. you can dissable the swap file, but then you'll need to either reboot every few days to clear the ram, or not use your computer for anything.

You shouldn't turn the swapfile off. Windows' VM system is specifically tuned to work with the assumptions that swap is available and there is at least as much swap as there is physical RAM.

Indeed, you'll probably find nearly all OSes today are faster and more reliable with swap than without it.

the ram dive gives you the speed you want of ram, but tricks windows into using it responsibly.

I do not believe your hypothetical situation would be better. Windows' swap algorithms are different from Linux (which is different again from FreeBSD), but I sincerely doubt it would work better with swap on a ramdrive than it would with simply more RAM.

(Note that I'm talking about Windows NT here, not Windows 9x.)

Plus i might want to use my entire (hypothetical) 2 gigs of ram and still have a ram drive to boot. and mayb

Wait 45 days before buying a new PC

[amigoro]

Why not wait 1 year, 2 years, 3 years, a decade?
After all, tomorrow never comes, especially in the Computing world.

Moderate this comment
Negative: Offtopic Flamebait Troll Redundant
Positive: Insightful Interesting Informative Funny

Re:Wait 45 days before buying a new PC

[sacremon]

The "wait 45 days" is likely in reference to the anticipated released of the 939-pin version of the Athlon 64 FX-53. The present 940 pin version requires registered RAM, which slows it down a bit. The 939-pin version will work with unregistered RAM, allowing it a boost in speed in many applications.

Initial observations

[Pidder]

A quick glance on the system setups shows that they have used RAM with almost the same CAS-latencies in all the setups. The AMD CPUs benefit from low CAS to a greater extent than the P4. When an Intel fanboy site like Tomshardware wants the p4 to beat the Athlon they usually use very slow ram on the Athlon setup, which is of course overlooked by most consumers.

Re:Initial observations

[ThaReetLad]

The other thing they do is rely heavily on synthetic benchmarks like PCMark. A little while ago the inquirer ran a story that pointed out that a heavily overclocked p4 running at over 4 GHz got impossible low scores on some synthetic benchmarks. This turned out to be because the windows internal clockspeed counter was just 32bits and there was an integer wraparound. What they failed to notice was that this meant that these so called benchmarks were baseing their scores on the reported clockspeed of the processor, and therefore must have favoured intel.

Re:Initial observations

[mikehoskins]

What I do notice it that they seem to me to be VERY chatty about Intels. When they talk about AMD's, they seem to be quieter.

I notice that an AMD could nuke an Intel on a test and get little press. When an Intel edges out an AMD, it gets more press.

Perhaps it's big news when Intels can keep up with AMD's, these days! :-) I would love to see 64 bit versions of the software in these comparisons, to show how future proofed the AMD's are.

And I know that everything I've said here is anecdotal and that I'm biased toward AMD chips....

Re:Initial observations

[drinkypoo]

I've certainly noticed that people are more or less willing to take for granted that two things are true about AMD processors. One, they do more per cycle. This should be clear, anyway, because they have more functional units and the functional units are more flexible. Also they have the credibility of having built a multitude of RISC designs over the last few years, most of which have had x86 emulators on the front of them of course. Well, and the back, it wouldn't do to fetch and never retire. Anyway Two, they are much cheaper than intel processors. Sadly intel outstripped AMD in terms of bus bandwidth some time ago and AMD is just now catching up again with the processors with integrated memory controllers - Since that is separate from the bandwidth used to the north bridge. It seems that HT should give about 1/2 the performance of the P4's FSB, but since it doesn't have to carry information from the CPU to main memory (FX-53 has a DDR 400 dual channel memory controller, which should be plenty of memory bandwidth for anyone. Of course DMA still has to occur via HT but in most cases this should not be a serious problem. (Using system memory for AGP textures will still be slow, though of course still faster than loading them from disk all the time.)

So, the hot AMD processor (FX-51) currently beating up on the hot intel processor. The FX-53 is even more destructive (about 10% faster still) and I doubt that it will be substantially more expensive than the P4 EE 3.2GHz 2MB cache, which is already defeated in the benchmarks by the FX-51.

So yes, with the release of the Hammer-core processors, it is unusual for intel to be able to keep up with AMD these days - As it was with the Athlon before it. Remember when the Athlon's double-pumped bus made it two or three times as fast (in terms of FSB) as the intel processors? And how intel processors had less cache, and typically slower cache? Since the release of the K6 intel has been running scared, even in spite of the K6's many flaws. The Athlon was the real sign that AMD was ready to compete with everyone, that's really an amazingly slick chip and there's a multiprocessor version, so AMD targeted basically every space below supercomputing with that processor, and had good success with sales nearly everywhere. (Actually the K6 sold quite a few units also.)

Re:Initial observations

[D. Book]

I'm an occasional reader of Tom's Hardware, and they always stressed the importance of fast memory timings in affecting overall system performance.

Until recently. Check this recent test, the conclusions of which are contrary to what they have previously said, and contrary to what many speed-freaks have come to believe due to constant repetition.

For those who can't be bothered reading, the bottom line is this: memory timings matter little in affecting overall performance of current platforms.

Heat

[shawkin]

With the case open, this thing runs at 178 degrees. In a practical sense, all the other benchmarks are less important.

It is not going to be easy to cool. It is not likely to be suitable for clustered processing. It is not likely to be particularly reliable.

This article illustrates the diminishing returns of the current Intel CPU architecture and processes. Soon, both AMD and Intel will be forced to explore new designs similar to the IBM Power 5.

Given the time, effort and money involved in developing a new CPU architecture, the near and medium term future may lie with IBM.

Re:Heat

[xeper]

With the case open, this thing runs at 178 degrees. In a practical sense, all the other benchmarks are less important.

Just wondering: Is the described setup with the case open & lying on its side actually better or worse for cooling?

With the case closed you have a nice airflow from the frontside fan [1] over the CPU-Cooler to the backside/PSU-Fan adding to overall cooling [2]... OTOH having the case open makes it less likely that the CPU-Cooler tries to cool the CPU with the already heated air from the case...

[1] If installed but since I noticed that HD temperatures decreased from >40C to ~20-25C just by the gentle blowing of a 7V fan over their cover all my boxes have them...
[2] Like in my two Via C3 powered, passive cooled boxes :)

--

Re:Heat

[Anonymous Coward]

RTFA! The case was cooled by having it open and on its side.

Re:Heat

[TheLink]

I guess it depends on how cold the ambient air is, and how many fans you have.

Even so, case open + 75 watt desk/floorstanding fan is quite hard to beat :). Get two and you have redundancy, plus the better quality fans are pretty quiet. But case open = easy for damaging stuff to get in.

I wonder if I could go passive cooling if I underlock my 2500xp Barton to 800MHz? Thing is the other fans make noise too (GPU, power supply, HD fans), so the actual reduction probably isn't worth it.

Re:Heat

[Anonymous Coward]

Soon, both AMD and Intel will be forced to explore new designs similar to the IBM Power 5.

POWER5 (dual-core, SMT) is targeted at the server market. Did you mean the PowerPC 970, i.e. a low-power, reasonably-clocked processor? Do Banias and Dothan ring a bell? There are rumors flying around that Intel could abandon NetBurst in favor of the Pentium M micro-architecture.

Re:Heat

[xeper]

RTFA! The case was cooled by having it open and on its side.

I RTFA. The case was open and on it's side during testing. However, the article didn't say if this setup was chosen for cooling reasons or just because it is easier to change CPUs with the case placed this way.

Re:Heat

[shotfeel]

I hope they didn't do it for cooling reasons because having the case open can cause some parts to run even hotter.

It can be the equivalent of opening a window in the middle of the winter to cool your house. If your furnace is still on, that room may end up cooler, but if the thermostat is in or near that room, some parts of the house could end up even hotter than they were before.

Also, most systems are designed for a specific air-flow pattern. If you open the case, you disrupt that flow. The fan can no longer cause cool air to flow over the "hot spots" like the CPU which can cause them to get even hotter.

Re:Heat

[Anonymous Coward]

Everytime I see someone say a processor is running at 100+ degrees I just laugh and remember my AMD Athlon XP, sitting there running at 25 degrees celsius.

Re:Heat

[igny]

CPU-Cooler tries to cool the CPU with the already heated air from the case...

when the cpu fan rushes the air through the radiator the hot air goes sideways, and then relatively quickly goes up, leaving the case. The colder air from the room displaces the warm air in the case.

Open case lying on its side is a very good setup to cool the cpu using ambient air. If the air in the room get circulated, say, by a ceiling fan or an air conditioner, the setup is as good as it gets.

Re:Heat

[Hoser McMoose]

Just wondering: Is the described setup with the case open & lying on its side actually better or worse for cooling?

The only answer to that question is: It depends.

Ideally your case will be designed to have a decent flow of cool air through the case, and in particular the cool air will flow over the processor while the warm air will be sucked out the back. This is better than having the case lying open and on it's side where the only flow of cool air will be from free convection (ie hot air rising).

In reality though, there are dozens of factors that can affect this. Certainly the number, speed and placement of fans in your system can change things a lot. Occassionally adding fans can actually make cooling WORSE! (don't tell the overclockers this though, they're doing a good job of keeping the economy going by buying 75+ fans for their PC). Also things inside a case like cables, add-in cards and hard drives can change the flow of air.

In the end, chances are that it's not going to make a big difference one way or the other. What might make a bigger difference is the method used to measure and report temperature, something that is notoriously inaccurate on most motherboards. Even with the on-die thermal diode of the P4 I wouldn't trust the temperature monitoring software to more than about 1 digit of percision.

Still, this chip is getting too hot. Intel states that the maximum temperature the chip should ever reach is only 73.2C.

Re:i havent read the article

[tttonyyy]

does it actually have nay insightfull comments, or things we havent heard of already or is it pretty much like the previous cpu review made:
This Just IN, NEW CPU FASTER THAN OLD CPU

Nothing exciting really. Summary is basically this: Amd for 3D games, Intel for MP3 and DivX encoding (and marginally for some scientific software). At the end of the day it all depends what software is running, so there's no clear way to define which is "better" for the masses.

Nothing changes there, I guess.

Question

[Saint Stephen]

Since a shorthand way of describing the benchmarks is (1) Athlon FX 64 has much higher (default) memory bandwidth than P4 3.2, (2) Athlon FX does much better in UT2003 than P4 3.2 (plus other results like P4 doing better in Rendering/Encoding), can I conclude that UT2003 is memory bound?

And thus since I have DDR400 in my P4 3.2 and can overclock it to to 220 FSB to get Sandra benchmarks of 5600 MB/sec (non OC/d it's like 4900), that I can expect to get similar UT2003/4 numbers as the Athlon FX? Obviously one could theoretically OC the FX to get even higher memory numbers.

Summary of the article

[heironymouscoward]

1. AMD64 is better for games
2. Intel Northwood P4 3.4 is good for general use.
3. Intel's new Prescott is too hot.
4. Whatever you buy will be redundant in 2 months.

Plus ca change, plus ca reste la meme chose.

Re:Summary of the article

[rokzy]

redundant? why, are they going to start giving everyone lots of free CPUs? or is all software suddenly going to become much more efficient so we don't need powerful CPUs any more?

just because you see a word lots of times on /. doesn't mean you know how to use it.

Re:Summary of the article

[Mr Guy]

[quote]just because you see a word lots of times on /. doesn't mean you know how to use it.[/quote]

Ironically, your examples do not give particularly convincing evidence you do either. Your second example is much more convincing for "obsolete". The first one is marginal. Perhaps you meant: redundant? why, is your processor going to reproduce? Will you suddenly have TWO or possibly THREE processors?

Re:Summary of the article

[rokzy]

redundant ( P ) Pronunciation Key (r-dndnt)
adj.

1. Exceeding what is necessary or natural; superfluous.
2. Needlessly wordy or repetitive in expression: a student paper filled with redundant phrases.
3. Of or relating to linguistic redundancy.
4. Chiefly British. Dismissed or laid off from work, as for being no longer needed.
5. Electronics. Of or involving redundancy in electronic equipment.
6. Of or involving redundancy in the transmission of messages.

look again:
lots of free CPUs means the one you bought is unnecessary. alternatively if software is more efficient, your new high-speed CPU is unnecessarily powerful.

get it now?

Re:Summary of the article

[mog007]

2. Intel Northwood P4 3.4 is good for general use. Woah there. For general purpose use I'd go with whatever would be most cost-effective, and AMD is still clobbering Intel in that reguard. The P4 would be great if you had a dedicated SETI@home server running 24/7 and you wanted to get some serious data analyzed, but not very many people would fall into that category, thus negating the "general" portion.

Re:Summary of the article

[Mr Guy]

Look again, lots of free CPU's mean the free ones are superfluous, not the original. Alternatively, if the software is more efficient, the CPU contains redundant power, the CPU itself is still required. No matter how efficient your software is, it still requires at least one CPU to run it.

Re:Summary of the article

[Krondor]

Just an update:

1.) AMD64 is better for games.
2.) Intel Northwood P4 3.4 is good for general use
3.) Intel's new Prescott is better then Northwood for general use/video encoding especially with SSE3 in the future, but it runs too hot.
4.) Wait 45 days for new mobo's with new sockets and PCI Express.

Re:Summary of the article

[heironymouscoward]

Redundant means kicked out of a job where I come from. "I was made redundant by a computer." Actually exactly the same as obsolete.

Redundant also means excess to requirements, which is a twist on the previous meaning. "Those cables are redundant."

Redundant also means _deliberately_ excess to requirements, which is yet another twist. "We use redundant cabling to ensure scalability."

So, "redundant" as a designed attribute of a system is almost purely opposed to "redundant" as an unfortunate circumstance of change. But both meanings are correct, and while you may think of RAID disks when you read "redundant", using it in a sense similar to "obsolete" is not incorrect, just original and somewhat poetic.

And, you have to admit, it provoked some interesting and stimulating discussion.

OK, DAMN YOU ALL, I JUST USED THE WRONG WORD!!!

Shit. Mondays are the worst.

Mods are lamers, Post Parent DOWN

[Anonymous Coward]

This is an inane and stupid comment. How on earth can it be modded up? I know that the average /.'er is lazy and easily taken in by mod trolls, but this is pure and simply a sign of ignorance and stupidity on the part of anyone modding the parent up. Sigh. Maybe the mods are still getting over their spring break hangovers?

Throw some G5s into the mix

[oingoboingo]

What I'd like to see in a huge multi-CPU benchmark like this are some Apple G5 systems thrown in too. Decent cross-platform tests are hard to find, but given OS X's UNIX underpinnings, it may be possible to come up with a set of tests that are run on x86 Linux and OS X which have an identical code base, and which do not artificiallly advantage one architecture over the other. One thing I've found since switching to OS X about 6 months ago...the Mac community still lacks a really good site which does solid, rigorous benchmarks of Mac hardware/software...and there are a lot of myths and misinformation doing the rounds on various Mac forums (as there are on PC forums too). A well controlled multi-CPU benchmark including some Macs could go a long way to alleviating this.

Re:Throw some G5s into the mix

[MacEnvy]

Try www.barefeats.com. They're pretty good about keeping their benchmarks clean, and they do some cross-platform benching (although they are primarliy a Mac site).

Re:Throw some G5s into the mix

[thatguywhoiam]

This site compares Macs to Macs... its sort of useful.

This site actually has a German G5 vs. Athlon benchmark posted right now.

Neither one is like Tom's (good or bad)... but its something.

Re:Throw some G5s into the mix

[ValourX]

Apple won't send G5s to people who intend to benchmark against other computers. I know because I've talked to their PR department. They don't want to lose any benchmark tests.

Next time I won't mention that part of my testing.

-Jem

Re:Throw some G5s into the mix

[Anonymous Coward]

What compiler would you use under OS X? What compiler would you use under Linux/i386? Intel's compiler is often better than gcc. Assume you pick icc. Are performance differences due to the operating system, the compiler, the architecture, or a combination of the three?

Re:Throw some G5s into the mix

[Nimey]

I can see one problem with your idea: compilers. Say that everything running on our machines must be compiled with the same rev of gcc. What if gcc on Platform A isn't as highly evolved or optimized as on Platform B?

Say that instead we use compilers made by the CPU manufacturers, since they presumably optimize the best. No, can't do that. AMD, for example, hasn't the resources of Intel and can't write their own.

Just being the devil's advocate.

Re:Throw some G5s into the mix

[Anonymous Coward]

I can see one problem with your idea: compilers. Say that everything running on our machines must be compiled with the same rev of gcc. What if gcc on Platform A isn't as highly evolved or optimized as on Platform B?

Say that instead we use compilers made by the CPU manufacturers, since they presumably optimize the best. No, can't do that. AMD, for example, hasn't the resources of Intel and can't write their own.

So what you do is you use the "standard" compiler for the platform, whatever that may be: gcc for Linux, MSVC++ for Windows. I don't know offhand what it is on Macs.

The point is that it's nonsensical to compare the G5 against the Athlon anyway. It's meaningless which has the better raw performance, because nobody's using them raw. The only useful question is, will my software run faster on a G5 Mac or on an Athlon + WinXP? And the differences in the compilers are going to contribute to the benchmark, not ruin it.

Re:Throw some G5s into the mix

[Sloppy]

The right thing to do, is just use the best compiler you can find for each platform. Use different versions of gcc if you have to, use Intel's compiler for the P4, etc. Use whatever gets you the most for each platform, because that's what an end user who has CPU-bound work to do, is going to do. AMD does have a disadvantage here, but it's a real disadvantage that you would see as a user, so it's relevant to the benchmark and perfectly fair.

I don't get why AMD hasn't been putting effort into compilers. I realize they don't have Intel's luxuries, but I'd think we're just talking about one or two salaries. They wouldn't have to write a compiler from from the ground up; they could work on gcc.

Re:Throw some G5s into the mix

[TwistedGreen]

It's easy to have positive Karma. It's more difficult to think before you post.

It's kind of a fun game to try to predict what the highest-modded posts will be for Slashdot articles before looking at the comments. It's an interesting bit of groupthink that goes on here... once you're acclimatized to the Slashdot mindset, you almost don't even have to read it to know what's being said.

Someone getting (+5, Insightful) for saying that "Apple does it too"? All too predictable.

Re:Throw some G5s into the mix

[Hoser McMoose]

How about you just use the fastest compiler for each chip? Or the compiler that will actually be used for running your code?

I know a lot of people like to complain about the effects of compilers in benchmarks (particularly SPEC CPU, the most widely used cross-platform benchmark), but compilers are a critical part of the equation for performance. So what if GCC is more optimized for Platform A vs. Platform B, the end result is that Platform A will run your code faster, and that's what really matters.

Pretty poor LinPack performance...

[nickovs]

Taking the opportunity for a moment to troll, flame bait and be an annoying Apple user, I think it's worth commenting how piss-poor the P4's LinPack performance is. The Apple Xserve G5 gets 4.5 Gigaflops out of each of it's two 2GHz G5 processor when running HPC Linpack, as opposed to the 3.4GHx P4 "Extreme Edition" which peaks at just 1.3 Gigaflops. Anyone looking to do serious scientific calculations rather than just playing Quake should not be using Intel hardware these days; it just doesn't keep up with the PPC G5 for floating point.

Re:Pretty poor LinPack performance...

[Anonymous Coward]

Not saying the G5 doesn't offer better FP performance, but you might try citing some evidence other than what's on Apple's site. They don't exactly have the best track record when it comes to impartiallity (No manufacturer does). Let's not forget "The worlds fastest personal computer" fiasco.

Re:Pretty poor LinPack performance...

[AmigaAvenger]

Don't forget the whole thing about the being the 'first 64 bit computer'

Re:Pretty poor LinPack performance...

[maswan]

Assuming you stay within the hand-tuned codes that are available. You won't see anything near that performance from compiled code, and that's with a good compiler. I've done some tests myself, and the G5 performes about the same clock-for-clock as the Opteron. And these days, the Opteron clocks a bit higher...

Re:Pretty poor LinPack performance...

[Professor J Frink]

And don't forget the whole thing about being the first to move their product base to 32bit RISC (where they were beaten to it by quite a few years by Acorn and their ARM-based RISC OS range).

The more you look at things the more you realise it's all been done before by someone else years ago.

Re:Pretty poor LinPack performance...

[jstott]

Taking the opportunity for a moment to troll, flame bait and be an annoying Apple user, I think it's worth commenting how piss-poor the P4's LinPack performance is.

The AltiVec processor on the G5's is a vector coprocessor. If your compiler/library is set up to use it, that's good for a 4-5x increase in floating-point speed. Essentially the CPU does a block of mathematical operations in parallel--Cray mainframes work the same way, only more so. This is different from pipelining in that it's a true parallel operation. I think the AltiVec can do vector integer operations as well, but that won't change the LinPack performance.

Note too that the boost from a vector processor only works on specific types of floating point operations, most notably matrix math, so it's not a magic cure-all. Also, the data has to be in the right format and loaded into appropriate registers, so it helps to have code written specifically to use vector operations (although a good optimizing compiler can still do a lot of the work for you)

.

-JS

Re:Pretty poor LinPack performance...

[Anonymous Coward]

" Let's not forget "The worlds fastest personal computer" fiasco" and "first 64 bit computer"

I think Apple was refering to "Personal" Computing as stated above and not Servers/Mainframes/Super computers.

Re:Pretty poor LinPack performance...

[afidel]

They're still wrong, Opteron came out months before the G5.

Re:Pretty poor LinPack performance...

[nelsonal]

I think Genentech found that the AltiVec engine worked great for some code they had to sequence DNA they were buying lots of those G4 servers. I know that apple's processors held their own against much higher clock rate in Golomb Ruler and RSA calcs, our friends old imac was generally about as good as an x86 running a 50% higher clock.

Re:Pretty poor LinPack performance...

[mosb1000]

"They're still wrong, Opteron came out months before the G5"

Correct me if I'm wrong, but didn't the MIPS R1000 and the POWER 4 come out years before the Opteron. And I was under the impression that 64 bit Cray supercomputers came out decades before that. I think what apple was trying to say is that theirs is the first 64 bit personal computer. It all depends where you draw the line between workstation and personal computer I guess.

The Opteron is now and has always been considered a sever/workstation processor. This is because it if large, expensive, and power hungry. You rarely see PC's with that chip. The power consumption/size/cost of the G5, on the other hand, puts is squarely in the realm of the PC.

Re:Pretty poor LinPack performance...

[Anonymous Coward]

Damage has offered to bench Apple equipment in the past, if only someone will donate the equipment to be benched (since, as someone mentioned in this subthread, Apple is not interested in cooperating).

Re:Pretty poor LinPack performance...

[Anonymous Coward]

First of all, if the PowerMac G5 isn't a "workstaiton", I don't know what is -- It's a high-spec box that costs twice as much as an average PC.

Second, there was a DEC Alpha PC sold back in the late 90s. And by "PC", I mean a cheap-assed $1000 desktop box that nobody would confuse with a "workstation".

Re:Pretty poor LinPack performance...

[Ruliz Galaxor]

for floating point maybe, but that's more because you're talking about a RISC. Both Intel and AMD use a CISC, so actually you're more or less 'flaming' and 'trolling' the x86 instruction set rather than the CPU itself.

sig(h)

Re:Pretty poor LinPack performance...

[shotfeel]

I think what apple was trying to say is that theirs is the first 64 bit personal computer.

That is what Apple said. It just makes for more "fun" when people leave out the word "personal".

IMO though, its just marketing. Everybody uses qualifiers in order to say their's is the best, fastest, cheapest... Marketers take advantage of the fact that some people will miss those qualifiers.

Re:Pretty poor LinPack performance...

[afidel]

Well considering the number of PC's that I have made with either Opteron or Athlon64's (same chip basically) I don't think Apple was right in any case. Sure AMD's marketing hype might be positioning the Opteron as a 'workstation' chip, and Apple is calling the Dual G5 workstation a 'Personal Computer' but it doesn't negate the fact that there are people using Opteron and kin for personal computer use and there are people who are using the G5 in professional workstation capacaties.

Re:Pretty poor LinPack performance...

[Anonymous Coward]

Actually, as someone who writes complex matrix operations on the G4 AltiVec for embedded applications, I can tell you: You're dreaming if you think you'll get vector ops anything close to promised.

Not that it's bad, and the G5 has a memory bus that brings the Altivec almost into a bus/cpu balance, but you'll still never see anything like 4x speed increase (the thoretical limit of doing FP operations 4 times as wide). Where you think 5x could occur, I don't know...

Finally, matrix math is extremely touchy with regards to precision, and the fact that the AltiVec is 32-bit IEEE throughout, rather than having most intermediate values in 64/80 bit FP, is a serious hit for those operations requiring precision.

Almost nothing LinPACK is designed for (finite element, time-series analysis, etc.) would you really want to use the AltiVec. (Yes, I've also done FEA work professionally at a national lab).

Final note to trolls and fanboys: I'm not claiming the P4 is better than the G5, G5 is still probably better, but some things needed clarification.

Eric

Re:Pretty poor LinPack performance...

[Hoser McMoose]

So a single processor Opteron system with a PCI bus and AGP slot costing $1500 was a "workstation", but a dual-processor G5 with PCI-X costing $3000 was a "personal computer". Clearly these lines are more than a bit fuzzy.

As for power consumption, the Opteron and the PowerPC 970 are nearly identical clock for clock. The 2.0GHz PPC 970 has a maximum power consumption of about 70W, same as the Opteron (though neither are particularly well documented in this regard). The new PPC 970FX is a lower powered part, consuming only 37W maximum power at 2.0GHz, but even that isn't far off the lower power Opteron 246HE chips, running at 2.0GHz and consuming a maximum of 55W. AMD also plans on releasing a 30W Opteron running at 2.0GHz in the not-too-distant future.

In short, both are good chips, both target similar markets and both have similar power consumption numbers. The PPC970 has a slight edge in terms of size and cost at only 58M transistors vs. the Opteron's 105M transistors, but a lot of the Opteron's extra transistors are the extra cache (~60M transistors for 1MB of cache vs. ~30M transistors for the 512KB of the PPC970s cache) and the extra I/O stuff integrated onto the chip. In some applications the PPC970 may end up being a better choice, in others the Opteron would be a better choice.

Next stop, how about GPU sockets.

[Anonymous Coward]

CPUs are fast enough, but graphics cards are still a long way from perfection. The biggest problem is that it is limited to being shoved onto a AGP slot which has all sorts of difficulties.

Hence I'd like to see motherboards that have a CPU and GPU socket. Imagine a 2Ghz Radeon, that would rock!

Re:Next stop, how about GPU sockets.

[ShavenYak]

That's not a half-bad idea. Stick a DVI-I port on the I/O panel, integrate a RAMDAC on the board, and have a socket for a GPU. You probably also want a dedicated memory slot for video RAM, although it should be able to share system RAM as a lower-cost option.

The hard part would be getting the chipset and GPU manufacturers to settle on a single specification for the interfaces. But it would really be a win-win situation; it could reduce the cost of new graphics hardware and increase the profits for the GPU manufacturers, by reducing the amouht of redundant hardware thrown away at each GPU upgrade cycle.

Nothing new to see here

[twbecker]

The A64 is the best gaming CPU out there right now, whereas Intel excels at video encoding, etc, etc. Did anyone get anything new out of this roundup? I'd say the point of all this is that anyone wanting to build themselves a new rig NEEDS to wait a month or so, unless they want to buy into a platform that is essentially guaranteed to be non-upgradeable.

You and the mods really should quit smoking crack

[Anonymous Coward]

M'kay?

Scale matters!

[IceFox]

After reading it I get the following:

1) If you are doing anything in Lightwave by all means don't use AMD's XP :) There must be some major tweak they are missing.

2) Encoding type work XP seems to be the best bang for the buck (right now)

3) I had a difficult time understanding the results because most of the graphs didn't have a scale to go by. Some of them like the games you could figure out that 500fps is twice as fast as the slowest at 250fps, but in either case you didn't care. With lame from the looks of it the slowest was still faster then what I could rip from cd (need to test, but just off the top of my head). Maybe on the larger scale for a particular test all of the cpu's are very close together, but in the view of close up it looks like one is _way_ faster.

4) With all of the tests there wasn't one compiler test :(

-Benjamin Meyer

Re:Scale matters!

[mczak]

addressing 4):
I've seen several sites using compiler benchmarks recently. Quick summary: The P4 Prescott is about the same speed as the P4 Northwood, if compiling isn't done in parallel both don't have the slightest luck against Athlon 64 (can't even beat Athlon XP). However, if something like make -j3 is used, the P4 are very close to the Athlon 64. Some link (Visual Studio, can't remember the other which used gcc): http://www20.tomshardware.com/cpu/20040318/athlon- fx53-28.html

Re:Scale matters!

[Hoser McMoose]

1) If you are doing anything in Lightwave by all means don't use AMD's XP :) There must be some major tweak they are missing.

It's called SSE2. Lightwave uses it, the AthlonXP doesn't support it. The chip also has the lowest memory bandwidth of any of these chips.

As for the scale thing, you're right. A lot of the processor performance they're seeing just doesn't matter much for todays applications. One could argue that they might help out for future applications that strain the processor a bit more, but it seems a bit pointless to spend a lot of money today on a CPU for software that might appear a year from now when a much faster CPU will be dirt-cheap.

4) With all of the tests there wasn't one compiler test :(

I'm sure that one of the multitude of other reviews out there did a compiler test or two. If all else fails, there is always SPEC CPU2000. One of the CINT sub-benches tests compiling performance with GCC. You have to do a bit of sorting by sub-test, but Ace's hardware have a nifty SPECmine tool that can help you in this regard.

TROLL ALERT

[Anonymous Coward]

parent is "click my links" troll

Disappointing...

[Damek]

What, no G5? They should be ashamed of themselves.

(laugh...)

Vector processor

[Colin Smith]

Well, so far you've made the case for a vector processor, or an add on like AltiVec. How's about making one for a faster CPU?

Re:Vector processor

[Short Circuit]

Well, x86 has vector operations in mmx, sse and sse2.

Rather than add on more extensions, you can crank up the performance of existing extensions by performing more operations per second.

Of course, with that gargauntuan pipeline, you're going to lose some of that performance. But not much, if you're doing data whose operations don't require the results of previous calculations.

don't forget...

[thatguywhoiam]

Amd for 3D games, Intel for MP3 and DivX encoding (and marginally for some scientific software)....

... G5 for grating cheese...

I kid, I kid.

You can use 'em for protein folding too.

wrong programs for

[Anonymous Coward]

benchmarks. Lame is way too slow compared to GoGo, DivX sucks compared to XVid, blah.

Whenever anyone benchmarks Intel vs. AMD, why is it they always use media encoders that are faster on Intel chips? /*redundant question*/

As a side note, quite amusing noone ever remarks 'XVid' is just 'DivX' backwards :p

Re:wrong programs for

[Ruliz Galaxor]

actually it's XviD instead of XVid :P

LaGrande?

[slux]

Has everyone already completely forgotten about LaGrande?

The tech sites certainly don't seem to be making much fuss about the fact that Prescott has this technology already in it. I wonder how they can be that unknowing of it. There was this big Extremetech article on LaGrande though.

Even on Slashdot no-one seems to be bringing it up these days. For me, the benchmarks aren't even worth looking at with the knowledge that these processors are the beginning of the DRM revolution. Seems they're being able to sneak the technology inside every PC just as they've planned it.

Still, sticking with AMD is going to be just a temporary measure. Is there any talk about integrating DRM into the PowerPC? If not, maybe the next motherboard upgrade could be a Pegasos or one could just go with a Mac.

Re:LaGrande?

[thatguywhoiam]

Even on Slashdot no-one seems to be bringing it up these days. For me, the benchmarks aren't even worth looking at with the knowledge that these processors are the beginning of the DRM revolution. Seems they're being able to sneak the technology inside every PC just as they've planned it.

You bring up an excellent point, and one that I wonder about.

At some point, the Slashdot/Ars/Tom's crowd and others who are a little more informed will identify the 'last great un-hobbled processor', i.e. the fastest thing you can buy before the Palladium/DRM stuff starts to become baked into the CPUs. Right now it looks like AMD and Intel will both be using some kind of Trusted (ahem) platform and BIOS. A lot of people will buy that processor(s) and then there will be a drop-off. As it is, not too many people get excited about the difference between 3 and 3.4 Ghz... ask yourself, which would you buy: a 'non-trusted' 3Ghz CPU or a DRM'd 3.4ghz?

Of course the unwashed masses will not know the difference (it makes the interweb go fastar!!!!111!!1)... but the alpha geeks are the ones who pay the premium for the latest gear so AMD/Intel may actually register a hit in sales.

Re:LaGrande?

[Anonymous Coward]

It's not like the features in the CPU are going to do anything until you run an operating system on it that makes use of them.

Quite possibly, the more PCs are sold that end up running Linux/*BSD/others, the safer we are from the possibility that CPU/motherboard/BIOS manufacturers would want to release machines that refuse to run a non-trusted OS.

The parent post is rant and biased.

[yudan]

You'd better look at the results from TomsHardware before starting to rant about it. They are clearly drawing the conclusion tht AMD is better than Intel. Do NOT bring your biased personal taste toward other websites up here!

Re:The parent post is rant and biased.

[ph4s3]

I second that. I decided to go with the AMD 64 architecture based on a THG article citing price/performance ratios. I don't think I would have come to that same conclusion if I'd been reading tripe from some intel fanboy site.

Re: Redundant vs Obsolete

[Anonymous Coward]

4. Whatever you buy will be redundant in 2 months.

Obsolete is the word you probably meant to use. The British use the word redundant to mean "no longer needed for a job and hence laid off." Unfortunately, people often misuse that definition in an attempt to imply that something is not good enough for a task. (Hint: Redundant actually means something exceeds the requirements for a task.) So using redundant to mean not good enough is like confusing the words priceless and worthless.

Numerical Computation Society

[Anonymous Coward]

The NCS is sponsored by IBM.

Active Cooling for the future

[stecoop]

We have come upon a point in processor technology that Active cooling other than passive radiant exchange with active air movement technology (IE cooling block and a noisy fan) will be exceeded very shortly. Running at 178 degrees these guys will require active cooling system such as water, refrigerant. But there is a third type of cooling technology that is micro channel cooling: http://www.cooligy.com/micro_channel_cooling.html I would like to see a rock solid active cooling system implemented and run as well as today's fans or even refrigerators. As more money is invested in the area then we will see more active cooling systems.

Because the reviews would have to wait 6mo/1yr.

[Anonymous Coward]

Until SOME of the software gets ported to the mac. At which point, besides mac fans, who would care? Who-hoo! Apple is proud to bring you Civ2, Half-life, Doom, and other highly anticipated games! Get your copies today! Only $99.99/ea! BF1942 scheduled to come out in fall of 2006! & on the hardware front, we're proud to present you with the Geforce MX! Gorgeous & FAST graphics for the best gaming experience!

Maybe

[Udo Schmitz]

I thought Intel was killing their label of chips by speeds...

Maybe the "E" doesn't stand for "Extreme", but "Estimated"?

Gratuitous?

[Anonymous Coward]

What a stupid way to describe the reviews. Of course they're free. It's not like you're paying to read the reviews.

Maybe you meant comprehensive or inclusive?

That' It. Moores Law is over!!!!!

[The Anointed]

When did the 2.0 GH Pentium come out, around August 2001. And now we're reviewing 3.4 GH Pentiums 2.5 years later? Dead!!! Long Live Moores Law.

Pitting 64bit, vs 32bit

[polyp2000]

It seems a little dubious to pit a 64 bit processor (Athlon64) against a 32bit one.

The Athlon64 does surprisingly well in many of the tests, especially when you note that in the majority of benchmarks it is only executing 32bit code. I bet we would see a different story if the Athlon64 was running at its best ability eg running 64bit apps on a 64bit os.

How difficult would it be to do some benchmarks comparing two identical linux distro's running on the same processor but one compiled for 32bit and the other compiled for 64bit. That might be an interesting comparison.

Nick

Re:Pitting 64bit, vs 32bit

[twbecker]

I don't see how it dubious at all. As much as I'd like to see a 32 vs 64 bit benchmark like you describe, you have to keep in mind that nearly all A64 owners are running 32 bit software, and that that processor was designed to provide competitive 32 bit performance. I'll be in the market for a new PC in a couple of months and I know that I'll be evaluating the A64 on it's 32 bit credentials, since that is all that's available for the forseeable future.

Re:Pitting 64bit, vs 32bit

[polyp2000]

Yes, I'd agree somewhat with what you are saying. Although I look at the A64 from a slightly different perspective. To a certain extent the intention of the chip (as I see it), is to provide a bridge to 64bit, driving consumers forward to making the "next computing step" without removing backward compatibilty with the 32bit world.

I see your point about the majority of users running 32bit applications, really I suppose its stereotypically a windows world.It is a valid , and real world, useful benchmarking review, no doubt about that.

I just think that since linux is 64bit ready (eg SuSe) we do have tools freely available to show us exactly what performance benefits can be squeezed out of these beauties. From a linux user's point of view looking to upgrade, the A64 is a much more interesting option because we can take advantage of its 64bit goodness now, as appose to whenever Windows Shorthorn comes out.

Such a review could also be instrumental in showing that in many area's linux is ahead of the competition.

Re:Pitting 64bit, vs 32bit

[Anonymous Coward]

Try going here for a little 64- to 32-bit benchmark

Re:Pitting 64bit, vs 32bit

[shotfeel]

Maybe I'm missing the point, but how would 64 bit apps help the Athlon 64? How is a 64 bit app any different? A 32 bit operation (add, mult...) doesn't take any more clock cycles the the equivalent 64 bit operation.

It also seems that if data structures are unnecessarily made 64 bits wide (using 64 bit integers when 32 or less would suffice), its going to slow down a 64 bit processor -takes more bandwidth to move around 64 bit numbers vs. 32 bit, plus not as many will fit in processor caches.

Not to pick on the parent, but I see a lot of posts that at least imply 64 bit apps will be faster, but I don't know that that's true unless the data being manipulated actually needs to be 64 bits wide (or can use >8 GB RAM).

So I'll ask again, am I missing something?

Re:Pitting 64bit, vs 32bit

[p3d0]

You're right, though not because of the 64-bit thing. Going 64 bit actually has a small performance penalty. But in 64-bit mode, Athlon64 has twice as many registers, and registers are the fastest storage there is (kind of like an L0 cache), so doubling that more than makes up for the 64-bit penalty.

Re:Pitting 64bit, vs 32bit

[polyp2000]

It depends on what exactly it is you are trying to do. For example a 64bit cpu is going to be considerably faster at computing an operation on two 64bit numbers.

If you are doing any kind of calculation intensive operations having 64 bit is going to increase speed.

There are many benefits such as Music applications eg audio channel mixing applications; with 64bit you can realistically increase the mixing headroom for sound channels;

Quick synopsis to explain the point;

Basic physics tells us that if you add two identical waves together you double the amplitude. When you try to mix two 16bit sounds together on a computer you must halve the amplitude of both (and then add together) in order to keep within the 16 bit resolution. The more samples you wish to mix together the more you must reduce the overall bit depth of each channel to keep from clipping. It is common for digital mixers to increase the bit depth of mixed sounds to 24 or even 32bit in order to retain bit depth and resolution as much as possible before final mix-down to normal 16bit. 64bit is surely going to be a major bonus in this area.

Im sure there are many other applications such as Video editing / mixing or 3D /Raytracing work which benefit enormously from 64bit.

AMD wins.

[Jexx Dragon]

Not only that, but Intel has confessed that 64-bit extensions now lie dormant in Prescott, ready to be turned on in future versions of the Pentium. This fall, Microsoft will deliver a 64-bit version of Windows, and both AMD and Intel processors will run it.

Why are the 64-bit extensions disabled? Linux comes in 64-bit now, which clearly means I'll be buying a Athlon 64 over an Intel. Then agian, maybe I'll just go with a four or eight processor Opteron based system. I here the 8088s are good this year too... If any still exist. Of course, I suppose the review caters to gamers who are stuck with Windows (or WineX, which would probably run pretty quick on a Athlon 64-FX).

Still, I think AMD is in the lead right now, they got the Athlon 64 and Athlon 64-FX out a few months ago and Intel still dosent have one, or... Well they do, but the disabled the 64-bit extensions. Can't forget that they have to use the AMD64 extensions as well.

Re:oooh....

[Beek]

Hatred of overclockers AND Alert? I salute you :-)

AMD64 testing

[ValourX]

Yet another review that doesn't test in 64-bit mode.

I don't know why this wasn't deemed Slashdot-worthy, but here's an excellent review of a P4 3.2E versus an Athlon 64 3200+ in both 32-bit *AND* 64-bit mode:

AMD64 vs. i386 in FreeBSD

-Jem

Re:AMD64 testing

[Anonymous Coward]

I don't know why this wasn't deemed Slashdot-worthy, but here's an excellent review of a P4 3.2E versus an Athlon 64 3200+ in both 32-bit *AND* 64-bit mode

Errr, is it your own review that you are calling "excellent"?

Re:AMD64 testing

[shotfeel]

And note that sometimes 32 bit mode is faster.

I'm not very familiar with the tests used, but it seems the only time 64 bit mode was significantly faster was in the Offenc tests. I'm still trying to understand why that is the case. The author states its because "there are twice as many general-purpose registers available", which I'll check into when I get a chance.

The other results show 32 bit mode to be either on par or significantly faster than 64 bit mode. For now I'm ignoring the "Synthetic Benchmarks" since the author states "Stream measures...although I don't trust it" and it sounds like the Ubench tests couldn't even be run to completion.

Re:AMD64 testing

[ingenuus]

Please forgive my ignorance, but it just struck me:
- Can (i386) 32-bit code run in (AMD64) 64-bit mode?
- Must we abandon all 32-bit code when we switch to 64-bit mode?

It would be awesome if we could get the best of both worlds simultaneously: smaller 32-bit programs in most cases, and 64-bit for those (fewer?) applications where that would be more effective. Intel would have to compete in performance with the best of both modes.

If so, then GCC should compile to 32-bit mode when that would yield faster code, no? Or is there a penalty for switching back and forth between 32-bit and 64-bit mode?

Otherwise, Athlon64 would seem more like two (mutually exclusive) CPUs in one rather than an extension of i386. In which case, is it even possible to have a backwards compatible 64-bit extension of the kind I am describing?

Thanks in advance.

Re:AMD64 testing

[ValourX]

The processor itself can run both 32-bit and 64-bit code, but FreeBSD cannot at present. Yesterday a rather messy hack was posted to the AMD64/FreeBSD list that adds a basic level of 32-bit compatibility, but I haven't tried it out yet. The problem is in the libs, not in the 32-bit program binaries, from what I understand.

You could run 32-bit and 64-bit binaries side by side on the same OS in theory, and eventually this will be possible. However the only advantage would be full backwards compatibility... 32-bit binaries will not get the same level of performance as 64-bit binaries, as seen in the oggenc tests. In most cases there isn't much that has to be changed to move 32-bit code to 64-bit with the AMD64 ISA.

-Jem

compiler comparison

[pwagland]

Hmm. Just had a quick browse of the article, and noticed something a little funny. In the Sphinx speech recognition test they compared all of the chips with both the microsoft and the Intel compiler. What was strange about it though was that for every AMD chip the Intel compiler was faster, by up to 4%. However, for 7 out of the 10 intel processors the microsoft compiler produced faster code than the intel compiler!

Bizzare eh?

Short version: Don't bother!

[Junks Jerzey]

Note: I'm not trolling, nor am I an AMD zealot.

Yes, you can't go by raw clockspeed alone, but in this case its close enough. In short, 3.4GHz P4 is THIRTEEN PERCENT faster in raw clockspeed than the 3.0GHz P4. The actual performance increase is less than that. At the same time, BOTH PRICE AND POWER DISSIPATION have gone up by MUCH MORE THAN THIRTEEN PERCENT.

Bottom line: This is a completely uninteresting processor at the current time.

best quote

[Anonymous Coward]

Under load, the Northwood P4 3.4GHz hit temperatures of 64 degrees C (148 degrees Fahreneheit). In the same conditions, the Prescott raced past Northwood's peak temperature on its way to a steady peak of 78 degrees C (or 173 degrees Fahrenheit).

Does anyone need to fry a steak?

Unbiased journalism

[pwagland]

Now, sure, we don't expect these people to be totally unbiased, but where did they pull this from?:

The Pentium 4 'E' is an absolute monster in workstation graphics, capturing the top spot in three of the six tests and tying for it in one more. In the other two, the Prescott 3.4GHz is second only to the Athlon 64 FX-53.

By the way, that test that it tied? It tied it with the Athlon 64 FX-53. But then I guess they wouldn't get their advertising budget if they said:

The Pentium 4 'E' and Athlon 64 FX-53 are roughly equal in workstation graphics, with the P4E winning three of the six tests the A64 FX-53 winning two, and they tied one test. Overall though there was less than 2% difference in any test.

Has anyone benchmarked Antivirus software?

[TheLink]

Antivirus software is as realworld as it comes. Is Athlon64/Opteron faster or is P4/PG+HT? My guess would be Opteron (based on AV stuff potentially having more branch misses), but I haven't done any tests.

Think AV scanner on a linux box to scan email/web passing through it. How many raw executables per sec?

Also, consider copying lots of files from one striped HD array to another striped HD array with AV enabled. If HDD still saturated - how much idle CPU left to do other stuff?

Another idea...

[burnttoy]

Could someone PLEASE conjure up a half decent Socket A CPU?! I mean one that doesn't require some huge cooling system... then I can make use of all those lovely Socket A integrated mobos with a nice passively cooled (silent) CPU... It doesn't have to be State of the art - I'm running a Duron 1.8Ghz here and audio and image editting are running just fine... just like they worked fine on my 500Mhz K7....

Re:Another idea...

[ShavenYak]

You can passively cool a Duron 750MHz. I found this out quite by accident one day when I opened up my case and found that the fan had quit working at some unknown time in the past. I imagine you could even go faster, if you were using a heatsink that was actually designed for passive cooling as opposed to an HSF with a dead fan.

Multithreading time in seconds?

[Tibor the Hun]

Does anyone have any specs on multithreaded rendering in seconds as opposed to score.
"Score" doesn't really say much to me.

Was hyperthreading enabled?

[mrm677]

I don't believe that the benchmarks utilized hyperthreading...which is reasonable given that most people are interested in existing single-threaded apps. However the nice thing about a HT-enabled processor is that your system is still extremely snappy while crunching that MPEG encoding.

Re:Was hyperthreading enabled?

[CrashPanic]

RTFA
The test clearly states that hyperthreading WAS enabled.

Re:Was hyperthreading enabled?

[mrm677]

Fine. Were multi-threaded benchmarks used? If not, the use of hyperthreading wouldn't be reflected in the benchmarks. The benefits wouldn't be measured.

Re:Was hyperthreading enabled?

[Hoser McMoose]

Again, RTFA! They stated several times where multithreaded benchmarks were being used.

Re: fast forward a little...

[Anonymous Coward]

Windows 2020 Minimum System Requirements (single partition "home user" license)

• 1 THz 128-bit Genuine Intel (or 100% compatible) Processor (3 THz recommended)
• 32 TBytes of Ram (128 TBytes recommended)
• 100 PBytes of free disk space
• 16 GByte video card

Optional:

• 10 Tbit wireless ethernet
• 200 GByte removeable disc (encrypted, 40 MByte/sec transfer)

3.4 / 3.0 = 1.1333..., a mere 13% increase

[mosel-saar-ruwer]

Besides this test being ridiculously comprehensive, did anybody else notice the stat differences between the P$ 3.0 Ghz - 3.4 Ghz?

3.4 - 3.0 = 0.4

is not so important here as is

3.4 / 3.0 = 1.1333...

which is a mere 13% increase.

I didn't bother to RTFA, so tell me: Did Intel achieve anywere near a 13% increase in performance? 10%? 6.5%? 5%?

Also, remember that early versions of the P4 were MUCH slower than the P3 at the same clock speeds [owing to all that increased pipelining]. And note also that the P3 continues to retain something of a cult status among hardcore skinflints; compare e.g. the PowerLeap PL-iP3T and PL-P3SMP. Finally, there is a persistent fascination with low-voltage, low-heat versions of the chip, which were originally marketed towards laptops, but are now drawing a lot of interest from embedded/blade/appliance manufacturers.

Uh, what?

[Anonymous Coward]

4. Whatever you buy will be redundant in 2 months.

Uh, how so? Will whatever chip you buy magically reproduce itself in two months so that you have two processors? I think the word you're looking for is 'obsolete,' and even then your statement is disingenuous.

I think his is more informative than yours

[Anonymous Coward]

His points out that the worst the P4 E does is 2nd, in any test and that it wins half the tests outright. This cannot be said about the A64 FX-53, as it comes in 3rd in two tests, falling to both the P4E 3.2 and P4E 3.4 in one test and falling to both the P4E 3.4 and P4EE 3.4 in another. Thus, his wording emphasizes the better results for the P4E 3.4 than for the A64 FX-53.

So I think the wording on that page is fine, although it is clearly far from the only wording you could use to describe the results.

As to the 2% thing, that simply wouldn't be any fun, would it? The performance gains in benchmarks are typically marginal, this reviewer is simplying following the same pattern as every other benchmarker by not calling attention to the fact.

When not to take a review site seriously...

[SomeGuyFromCA]

The reference to "enough microarchitectural tweaks to kill a horse" was bad enough, but now this:

We start with memory performance, because these benchmarks are synthetic [...] and not always indicative of real-world performance. They [...], however, [...] present the opportunity to make all sorts of colorful graphs.

Re:When not to take a review site seriously...

[shotfeel]

Actually, I tend to take the results more seiously when the tester demonstrates they understand the limitations of the tests used and lays it on the line.

Re:When not to take a review site seriously...

[SomeGuyFromCA]

Yes, but if they know the test is useless, I'd rather they not do it just for the chance to make pretty graphs and add another page to their review.

Oh great

[Anonymous Coward]

CPU is getting faster and faster but broadband speed is still the same. I think my gigahertz CPU is sitting idle 99.999% of the time waiting for the Network. It's like having a race-car and have to do stop and go in 15MPH traffic.

Look ma - my CPU is output more heat than yours

Benchmarks

[jak163]

The differences between these processors are small compared to the differences between these and the next generation of processors. That's why I just go by price myself, and I go for an order of magnitude improvement with each upgrade. PC --> 486 --> 686

OK, but...

[brucmack]

...consider the price difference between those two processors. Here's one case where the AMD is the overly expensive part.

When has this NOT been true?

[brucmack]

This is how it has always been... If one buys the latest and greatest part, it's going to come with a large premium. And that price premium isn't going to correlate with the increase in clock speed.

Heck, the P4EE and Athlon 64 FX processors aren't even at the top of all of the tests, yet they cost how much more?

This may be an uninteresting processor to BUY at the current time, but it is my opinion that this is a very interesting processor to STUDY. With the Prescott, Intel increased the pipeline stages by roughly 50% and yet the Prescott processors are still competitive. Further, tests like this allow one to judge how well the new model of P4 scale with clock speed compared to the Northwoods.

In short, these processors are the future. Just because it would be silly to buy one now doesn't mean that the benchmark is useless.

Relatively Old-School Athlon XP?

[unborracho]

I call BS. If the Athlon XP is old school than my 1.4 ghz Athlon T-Bird is Ancient Lore =)

Re:oooh....

[Anonymous Coward]

Shut up, Tom, you're making us Canadians look stupid.

Re:oooh....

[tomstdenis]

First off, I'm not Canadian.

Second, it takes a big shot cannuck bastard to post as an AC... ooh big boy, too afraid to be associated with retard posts?

Tom

power pc

[name773]

any g5 benchmarks in there?

PCI Express - External Graphics Processing

[meehawl]

Replacement for both obsolete 32bit 33MHz PCI and AGP. Try fitting two top of the line (AGP) graphics cards on one motherboard.

You're right, but PCI Express goes even further than this. PCI Express point-to-point connections can run around 5 metres. So you can have external PCI Express connectors and so stack your graphics cards outside your CPU box.

Given the increasingly hot and power-hungry graphics card GPUs, perhaps an external, self-powered, self-cooled graphics box(s) might become mandatory for super-high-end gaming rigs?

Re:oooh....

[Anonymous Coward]

tomstdenis@iahu.ca

Multimedia editing needs the faster speed.

[MtViewGuy]

Here's where the extra speed of the CPU becomes useful: editing multimedia files.

With the proliferation of digital still cameras storing large-sized picture files and MiniDV/MicroDV digital camcorders where you can copy the video recording in digital form to the computer for editing, there is now a serious need for faster and faster CPU's to edit and process these multimedia files at a reasonable speed. Even today's so-called mid-range AMD Athlon XP 2400+ CPU is getting somewhat hard-pressed to do such work even if you have an ATA-100 interface 7200 RPM hard drive and 1 GB of system RAM installed.

Occupation Inside

[Anonymous Coward]

there is at least one reason not to buy products from intel....

intel built a cpu factory (fab) on occupied palestinian land. they help to financially support an oppressive and racist regime.

Go AMD!

http://www.inminds.co.uk/boycott-intel.html

Different Linpack code!

[Hoser McMoose]

You're comparing Apples and oranges here! (no pun intended... honest! :> ).

The Linpack code used in this test was really designed to demonstrate the memory subsystem characteristics of P4 vs. the Athlon, not to crunch data. This should be blatently obvious even to someone making a troll/flamebait/annoying Apple user post, since the benchmarks you quoted show 3.2GHz Xeon processors (nearly identical to the P4) crunching at up to 4.35GFlops.

A more accurate view of the P4's capabilities for scientific computing is listed later in the review. Specifically the ScienceMark BLAS DGEMM tests is a pretty close approximation to the Linpack results that Apple is reporting. These tests show the 3.4GHz P4 'E' maxing out at ~4.1GFlops. Not quite where the G5 is, but pretty close. The G5 ends up being a fair bit faster because this sort of matrix solving is nothing but double precision floating point multiplies and adds. The G5 has this nifty FP Multiply-Add instruction that does both of these instructions at the same time, and the chip can do two such instructions per clock cycle. The P4 lacks this Multiply-Add instruction, so it needs to use two separate instructions, each with a 1 cycle throughput. The SSE2 unit allows each of those instructions to operate on two bits of data at once, so the total works out to the P4 having half the theoretical GFlops per clock cycle that a G5 has. In practice the P4 does a little better than this because it comes closer to matching it's theoretical peak than the G5 does (probably a memory subsystem issue).

side note: Altivec could potentially offer very high performance for this test, but it does not support double precision floating point numbers, only single percision, so it won't cut it.

Cost-benefit curves (performace)

[jago25_98]

With technology it's the performance you get for the bucks that is most easily compared.

Where can I find a constantly updated cost-performance curve for:
- CPU's
- graphics cards
- HDs

Basically you might have performance on the Y axis and cost on the X axis. I'd draw a nice diagram if slashdot would let me...

|
Performance |
|
-----------
Price

Of course, manufacterers also know that the best situation for them is to have unique technologies bound to certain software that cannot therefore be compared, thus outing them from competition.

intel old news

[Ogman]

Anything about intel is old news. All intel and microsoft do anymore is talk about what they are "going to do", which mostly seems to be DRMing us to death. AMD and Linux are actually advancing computing technology. Intel and Microsoft are starting to remind me of SCO, just putting out news releases to remind us all that they exist.