Xeon vs. Opteron Performance Benchmarks

QuickSand writes "Anand got his hands on some of Intel and AMD's enterprise processors including 4MB L3 Xeons, and put them to the test. Results were a little varied as 4-way Opteron systems seemed to fare the best, although dual Xeon configurations almost always beat dual Opterons. The exact benchmarks are here."

Comments

xeon vs opteron?

[Anonymous Coward]

How about Athlon? How Ghz is better, right?

Re:xeon vs opteron?

[Anonymous Coward]

Pfft. An Apple 2GHz dual PPC is 4,537% faster, STEVE JOBS NEVER LIES!

Re:xeon vs opteron?

[Directrix1]

Since the Opteron costs less than half of the Xeon, then Opteron wins right?

IA-32e vs IA-32

[Stonent1]

Can somebody tell me if the IA-32e processors will be in the socket 478 format to work with existing boards, or will they require a whole new socket and chipset (rather than a bios update) If they really are just "extensions" then I don't see why anything special would need to be on the motherboard correct? The cpu should switch into 64bit mode whenever the OS tells it to right?

Re:IA-32e vs IA-32

[Anonymous Coward]

No desktop versions will use new socket-775.

Re:IA-32e vs IA-32

[irokitt]

I don't really know, but I think Intel's 64-bit chips will probably use a Tejas-style clip system, not pins. Technology that won't work in a current motherboard. But, once AMD upgrades to socket 939 for the FX-51, it won't work in current boards either. AMD and Intel are both set to release the new sockets at about the same time PCI-Express comes out, so upgrade-happy people will need to buy new motherboards anyway.

Re:IA-32e vs IA-32

[adler187]

Technically since the new Prescott chips have the instructions on them, yes you could buy a 478 Prescott. You would just have to hack the chip to activate the instructions (good luck with that!).

As for a real IA-32e chip with the instructions enabled, Intel has stated that they arent coming out for while, and since Intel is moving the P4 to the new 775 chipset in a few months, I would bet that they would also be released under this new chipset. Heck, they might not even release their IA-32e chips within the lifetime of the 775 chipset.

Personally my Athlon 2600 just fine enough for me.

Re:IA-32e vs IA-32

[Anonymous Coward]

Intel has said that the initial 64-bit chips will be Xeon server processors -- which don't currently use Socket 478 anyway.

Re:IA-32e vs IA-32

[Anonymous Coward]

need to buy new motherboards

This makes me want to throw up. The last motherboard purchase I made, it was a chore finding one with the _least_ amount of features. Need an AMR riser slot? Fuck no, I'd rather have the PCI slot back. Need integrated sound? No, integrated sound makes my already bad speakers sound worse. It must've been tough figuring out how to make a decade's worth of improvements in technology amount to nothing. I have an ISA soundblaster from 10 years ago that sounds better than the onboard sound on my last motherboard. Need integrated video? I won't begrudge you this. Some people build clusters with their motherboards, and a video card is needed to boot, but if I have a choice I won't buy a mobo with integrated video.

In short, I want a motherboard with slots for RAM, an AGP slot, a socket/slot/hole for a CPU, PS/2 hookups, serial and USB connectors, and the rest of the board filled up with PCI (or PCI express) slots. That's the ticket.

Re:IA-32e vs IA-32

[starbuck8968]

Naconas are socket 603 like Xeons. I have no idea when desktop Prescott's will see the AMD-64 stuff, probably when LGA774 comes out at the soonest.

Re:IA-32e vs IA-32

[Octorian]

Well, any basic standardized and commoditized integrated I/O is a good thing. I'd also be fine with on-board Ethernet (good chipset, of course) and on-board SCSI/FC/SATA/etc. But yeah, I'd rather add my own cards for video/sound.

Re:IA-32e vs IA-32

[Loki_1929]

" Can somebody tell me if the IA-32e processors will be in the socket 478 format to work with existing boards, or will they require a whole new socket and chipset (rather than a bios update)"

They're disabled in all socket 478 chips. The new Pentium 4e chips (Prescott core) supposedly have the extensions, but they remain disabled. Technically, it may be possible to gain access to those instructions through some sort of BIOS hack, but you would also need to use an Operating system that can detect, support, and make use of those new instructions. Also, you risk using unfinished or untested parts of the CPU if you do manage to gain access and use the extensions. There would be no benefit other than simple tinkering.

"If they really are just "extensions" then I don't see why anything special would need to be on the motherboard correct?"

You still need a CPU that supports the instructions, and which has them enabled. Technically, if Intel released Prescotts in S478 form with IA-32e enabled, it should work fine with an existing motherboard which would otherwise support the Prescott chip you're using. The probablility of Intel taking the time and effort to do this less than a quarter away from a whole new socket is virtually nil.

"The cpu should switch into 64bit mode whenever the OS tells it to right?"

That's not entirely accurate. Technically, what happens under AMD64 (the basis for IA-32e), is that specific instructions can be sent to the CPU to have it run code in what's called "Compatibility mode", which essentially allows it to behave as though it were a 32-bit CPU. The difference is that you're not 'switching' to 64-bit mode. You're either in 64-bit mode with the option for compatibility mode when needed (meaning you need a 64-bit capable OS), or you're in 32-bit and you're stuck in 32-bit.

If you're looking for 64-bitness, you may simply want to get an Athlon64. If you're waiting for 64-bitness on the Intel side of things, you'll be waiting until some time towards the end of this year. Good luck.

Re:IA-32e vs IA-32

[LordHunter317]

This makes me want to throw up. The last motherboard purchase I made, it was a chore finding one with the _least_ amount of features. Need an AMR riser slot? Fuck no, I'd rather have the PCI slot back.
You do realize it costs much less to put a AMR or CNR slot on a motherboard then a PCI slot right?

Need integrated sound? No, integrated sound makes my already bad speakers sound worse. It must've been tough figuring out how to make a decade's worth of improvements in technology amount to nothing. I have an ISA soundblaster from 10 years ago that sounds better than the onboard sound on my last motherboard.
Now its obvious you're trolling. Say what you want about AC97-based onboard sound (which nearly everythign is), but its good enough. Your speakers are much more likely the problem. The long and short is that all PC sounds cards equally suck until you get to professional grade gear.

Need integrated video? I won't begrudge you this. Some people build clusters with their motherboards, and a video card is needed to boot, but if I have a choice I won't buy a mobo with integrated video.
What does having this cost you? Its not like you have to use it, or that boards with onboard video cost signifcantly more.

Re:IA-32e vs IA-32

[shepd]

The thing is, it all comes down to demand and volume.

Example: Motherboard without "extras", 100,000 sales. Manufacturer's cost? I'd put it at about $30.

Motherboard with "extras", 10,000,000 sales. Manufacturer's cost? Probably about $31, due to the higher volume driving down bulk costs.

What this all means is that by the time the boards get to the consumer, there's a $5 - $10 price difference, and it takes an extra week to get the "special" non-integrated board. Oh, and you have to pay shipping + PITA charges on that board. Which makes it cost $2 *more* than the other one.

Basically, if you don't want the features, they can be turned off easily enough. Buy the board with them, it'll have higher resale value, and, if your PCI card manages to bite the dust, you have the onboard backup to tide you over.

Oh, and if you're wondering why the integrated board sell so much better, it's your fault! Consumers are demanding to get a PC for under $300. That means cutting corners and cutting down on parts.

Speaking of cutting corners, have any of you enjoyed ASUS's "X series" motherboards yet? The PCB substrate on them seems worryingly thin.

Re:IA-32e vs IA-32

[Yokaze]

> Now its obvious you're trolling. [...] The long and short is that all PC sounds cards equally suck until you get to professional grade gear.

You must be joking. Most of the integrated sound I've had the joy to listen to produced noticeable background-noise. The most obvious one was the Eden-M board. Most mb-producers don't give much about seperating the analogue part from the digital, so accompanied with an intergrated graphics card, you can practically hear how a window is restored. It is usually not the quality of the on-board sound, which sells it, but purely the capability of producing some sound.

Re:IA-32e vs IA-32

[Anonymous Coward]

A-bit, Asus, Tyan.. get a name brand motherboard. You're comparing bottom of the barrel mini-ATX onboard sound with a name brand industry leader's sound card (Creative Labs).

Compare apples to apples next time.

Get an nForce or nForce2 chipset board. I know quite a few people who own and love them. Myself included.

Re:IA-32e vs IA-32

[mindstrm]

"Now its obvious you're trolling. Say what you want about AC97-based onboard sound (which nearly everythign is), but its good enough. Your speakers are much more likely the problem. The long and short is that all PC sounds cards equally suck until you get to professional grade gear."

That is absolute BS.. I'm sorry. Plenty of on and offboard common sound cards and chipsets (all of which are ac97-based) sound like shit. Some, however, sound surprisingly good. Some older, ancient cards still DO sound better.. better dacs, better shielding, etc.

How is it my speakers are the problem, when I try both soundcards side by side, and the old one sounds much better, and closer to reference?

For that matter, my headphones were about $900... if they are the problem, i'll eat them.

Re:IA-32e vs IA-32

[The Spoonman]

For that matter, my headphones were about $900... if they are the problem, i'll eat them.

Okay, not a troll, but an elitist. You're one of the very few people who would spend $900 on a pair of headphones because MOST people can't tell the difference between them an the $10 ones they get at Sears. The world doesn't work to cater to individuals.

Re:IA-32e vs IA-32

[mindstrm]

Not an elitist at all.. I realize that my tatste in headphones is eccentric. I also happen to use various other headphones nowhere near as expensive.

The average person can easily tell the difference between $10 headphones and, say, a good pair of $50 Grado Sr-60 headphones... far from elitist.

I'm only countering the argument that "all sound cards sound good". They do not, on anything above the run of the mill cheapest consumer gear.

If all you want is video games and some radio tunes, of course you won't notice, if you are actually listening to real music, you will.

Re:IA-32e vs IA-32

[DJStealth]

They will most likely require entirely new sockets as they are 64bit chips, as opposed to 32 bit.

The 32e means its an extended version of the machine/assebly code modified from IA32 to work on 64bit processors and still have backwards compatibility.

Re:IA-32e vs IA-32

[Carnildo]

Need integrated video? I won't begrudge you this. Some people build clusters with their motherboards, and a video card is needed to boot, but if I have a choice I won't buy a mobo with integrated video.

My A7V333 mainboard boots just fine without any video.

Re:IA-32e vs IA-32

[Anonymous Coward]

Sounds like someone either needs to go buy a Mac or go buy a Dell. Either way stop complaining about actually having a choice with regards to your technology. People who don't like choice deserve Windows.

Re:IA-32e vs IA-32

[Anonymous Coward]

The fact that you're listening to an ISA Soundblaster through a pair of $900 headphones indicate that you have some problems in your head that no motherboard can solve for you.

If you really want to use your computer to listen to music or whatnot, get a VIA system with a seperate power supply, no fans and a flash disc (no moving parts), a soundcard with good digital out, a good D/A converter, and a nice tubed headphone amp, and then start talking to us about how AC'97 sound sucks and that your Soundblaster is better.

The jump from what you currently call good to this will be the same or bigger than the difference between Ac'97 and an ISA Soundblaster.

Re:IA-32e vs IA-32

[Moraelin]

You must be joking about AC97 sound, right?

In addition to what was already said about SNR (signal-to-noise ratio), integrated sound traditionally comes with _the_ crappiest, buggiest, most broken drivers.

It's the _norm_, not the exception, that, for example, if you have WinAmp started, you can't get any error sounds, nor the sound notification that Outlook just received an e-mail for you. In truly pathologic cases, an application might even hang up until everyone else stops trying to use the sound.

Also on my NT workstation at work, if I have WinAmp started (not playing, just started!) when I try to shutdown the computer, chances are 50-50 it will just hang up for 15 minutes, then throw a barrage of error messages at me. Go figure.

So, honestly, until the mobo manufacturers stop putting dysfunctional sound chips with dysfunctional drivers on their mobos, I wish they stopped making me pay for that crap.

Re:IA-32e vs IA-32

[Anonymous Coward]

Either way stop complaining about actually having a choice with regards to your technology.

If I had a choice, I'd have bought a motherboard without integrated sound. They don't make boards without lots of crappy extras integrated. That's no choice.

Pffft. These Intel vs. AMD flamewars are pointless

[Can it run Linux]

because EVERYONE knows that Intel always wins.

xeons/opterons market share

[chef_raekwon]

dual xeons have owned the market for a long time...it will be difficult (although not impossible) for AMD to topple this.

many people did not upgrade to Intel's Itanium, but rather were upgrading to their high end dualie xeon systems -- they run very reliably, and very fast. a few instances where we've put in dual 2.x ghz xeons for web/mail servers...and only a slashdotting could bring them down...(well, an exaggeration...but you get the point).

Re:xeons/opterons market share

[ePhil_One]

many people did not upgrade to Intel's Itanium

Folks were avoiding the Itanium because it was a disaster; slow and expensive. We've been looking at 64 bit computing for a while, because of the seamless > 4GB RAM capabilities. Intel's PAE extensions are OK, but they really didn't solve any of the problems we were having.

The net result was we went to 64 bit PPC architecture 3 years ago on those critical systems, And everything has been fine. AIX works great, and IBM's embrace of GNU/Linux means an easy learning curve for us Linux users.

Re:xeons/opterons market share

[hackstraw]

I have not heard that the Itaniums were slow since Intel came out with the Itanium2. Yes, the Itanium1's were dog slow. I've got 65 Itanium2 processors downstairs, and I'm happy with them. For our purposes (crunching numbers on very large datasets) the Itanium2 was the platform of choice because of its 64bit addressing, high memory bandwidth and good processor speed.

I wish we could get by with cheap Xeons, but they just don't cut the mustard for our applications.

Re:xeons/opterons market share

[flaming-opus]

Itaniums are expensive, but not outrageous compared to other high-end processors like Power4 or ultrasparc. They also perform quite well. They are definately better performers than xeons for most of our apps.

The problem with itanium is not that they aren't a good technology, but rather that intel is trying to shove them into the high-end of the market, which is a difficult place to compete. sparc, power, pa-risc, alpha have all been around for years, have established customer bases, and lots of businesses have invested tons of money in running them. It's a difficult place to introduce any new products.

Intel has been stymied trying to sell ia64 into that space, and has undercut itself, by continuing to improve xeon, which performs pretty well and is comparatively inexpensive. Most segments are going to migrate to the all-american mantra of "GOOD-ENOUGH, and CHEAP!" which describes xeons/opterons perfectly. The market segments that won't migrate in that direction are willing to pay the big bucks for stability, and reliability. They are very slow movers. Intel might sell some itaniums to these customers, but they'd better be willing to wait a long time.

I think a lot of people judge itanium by the yardstick of Xeon, and maybe should not. If itanium ends up simply as a replacement for pa-risc, alpha, and MIPS in the SGI and HP portpholio, that may be a success by some measures.

Re:xeons/opterons market share

[Loki_1929]

The Itanium2s and up are pretty decent so long as you're working with code designed for 64-bit/EPIC. Where you run into problems is with 32-bit code, or pretty much any code not designed/optimized for EPIC. There's nothing wrong with Itanium in-and-of itself; it's just not cut out for compatibility the way x86 is. Had Intel stuck with the original plan for IA-64 (which was to replace x86 from top to bottom), this would have been fine. You simply would have lost the ability to use old applications, but new ones would run reasonably fast. 10 years later, Itanium has its niche, does quite well within that niche, and sucks for everything else. :)

"I wish we could get by with cheap Xeons, but they just don't cut the mustard for our applications."

This is exactly why Opteron DOES compete with Itanium - if only indirectly. Opteron will never hit the big-tin niche, simply because it was never designed, nor intended to do so. What Opteron does is bring 64-bitness, and all the benefits therein to the mid-range crowd. This forces Intel to choose between giving up on Itanium as anything other than a big-tin chip, or losing half its mid-range customers to AMD. Losing such a lucrative market would be far worse for Intel in the long run than losing the 10 years of R&D sunk into Itanium, so they've chosen to bring the Xeon line to the 64-bit world. With the new Potomac core (Q1/H1 '05), the XeonMP will be the CPU of choice for Intelphile mid-range customers in need of Itanium's benefits, but conscious of cost. The result will be that Itanium's legs will finally be completely taken out from under it, and it will be resigned to little more than a handfull of extremely high-end big-tin servers each year.

Does this mean Intel should continue to develop Itanium, even if it becomes clear it can no longer sustain its own R&D? I don't know - I think that's a question for Intel's board to answer. What I do know is that AMD had it right in '98/'99 when they decided to help transition people to 64-bit CPUs without losing x86's incredible compatibility. The bottom line is that someone like you would have gladly gone with either Opterons or Xeons had the choice been given to you. Unfortunately for Intel's margins, you and those in your position now have that choice.

Re:xeons/opterons market share

[Glock27]

Opteron will never hit the big-tin niche, simply because it was never designed, nor intended to do so.

Heh, I guess the Cray Red Storm system kind of shoots down that theory... ;-)

Actually the design of Opteron beats Itanium for HPC, and the relative number of Opteron vs. Itanium HPC design wins bears that out nicely.

Re:xeons/opterons market share

[fitten]

Where you run into problems is with 32-bit code, or pretty much any code not designed/optimized for EPIC.

I doubt anyone would have ever thought to buy an Itanium system to run x86 code.

In any case, x86 compatibility is the golden cage of computing. It's horrible so everyone wants to get away from it but at the same time, there is so much software that people "can't live without" that they cannot abandon the stuff. People bash the x86 architecture and at the same time, bash anything that isn't x86.

The AMD solution doesn't do away with x86, it keeps it alive and will require x86 to continue to live on, unfortunately.

Re:xeons/opterons market share

[Loki_1929]

"Heh, I guess the Cray Red Storm system kind of shoots down that theory... ;-)"

Not really - I mean, I could design a cluster with a few million Pentium Pro CPUs and have it compete with the upper end Itanium boxes. Does that mean the Pentium Pro was designed to compete with Itanium?

The Opteron makes an excellent solution in many different scenarios, but don't take its flexibility to mean that it was meant to be in direct competition with Itanium. As I stated, however, it indirectly competes with (vis-a-vis Xeon) Itanium, and in fact, creates a situation in which Itanium cannot possibly be self-sustaining within a year from now.

Re:xeons/opterons market share

[Loki_1929]

"People bash the x86 architecture and at the same time, bash anything that isn't x86."

Well, I think that people look at the x86 architecture, and they can see the many, many horrible hacks that have been used to sustain it. That much is pretty obvious if you spend even 10 minutes looking over things. You sit there scratching your head and going, "What the hell? Why'd they do that?", and then realize it's because something, somewhere, was broken until they did it. The reason people don't like to start looking into replacement architectures is exactly as you expressed; the must-have software. You can try running that software under emulation, but the best architecture in the world is always going to take a performance nosedive when running code under emulation. I can look at what IBM has been doing, or even at what Intel was doing with EPIC back in the day, and I can say, "wow, that's pretty cool". But what I can't do is put down the x86, toss all the old software, and hope that all the new software, written for a completely new architecture, is going to work in some sort of reliable fashion. What you really get with x86 is 20 years of experience, and thus, a measure of predictability. In essence, you're paying for predictable problems (much better than unpredictable ones) with old, poor architecture.

"The AMD solution doesn't do away with x86"

AMD64 actually does get rid of a lot of garbage in x86 that is no longer in use. Take a look at the presentation (link at Ace's) by the guy who designed AMD64. He was actually pretty thrilled (well, as thrilled as this guy gets) about being able to dump a lot of the cruft x86 has accumulated. Unfortunately, many things had to remain intact, for the obvious reason of compatibility. I have to warn you though, the guy from the AMD presentation is a real ball of fire. (Although, the ex-Intel guy from the other presenation was pretty interesting and funny)

Re:xeons/opterons market share

[fitten]

AMD64 actually does get rid of a lot of garbage in x86 that is no longer in use

Thanks for confirming this. I haven't had time to investigate this as deep as I wanted... I was ***so*** hoping that when the "go-into-64-bit-mode" bit was set, that it really became a whole different CPU, and that CPU looks like a traditional load/store (some may call it RISC) machine with 16 GPRs. When the "go-into-x86-mode" bit was set, it would then look like an x86.

Re:xeons/opterons market share

[Shinobi]

Except that theory takes a severe beating. Compare entries 4, 5 and 6 here at Top 500, especially the number of CPU's. Notably, entries 4 and 6 use the same kind of high-speed interconnects between the nodes, so the difference can't be blamed on that.

The problem with the AMD approach is that you get the NUMA drawbacks not only against other nodes, but internally on the node. If the data CPU 1 needs isn't in it's own memory banks, it's got to request them from CPU 2, 3 or 4, with a latency penalty(Letting the memory controllers read/write freely from each other's banks doesn't sound like a good idea, really). It works well with databases, serving websites, file servers etc, where the processes don't need to share memory, talk to each other a lot etc, but for physics and chemistry simulations etc, you get some penalties

Re:xeons/opterons market share

[andydread]

Intel "period" has owned the market for a long time. Intel is a marketing powerhouse. They could sell cheese to a dog if the wanted to. The xeon is inferior to the opteron in many ways. eg. the xeon bus is comparable to a hub while the opteron bus is comparable to a switch. If u read the article and the summary u see that the testers chose opteron over the xeon. Why? Well for starters the opterons tested are half the price of a 2MB cache xeon. the xeons used in the article are 4MB cache which is going to be waaaay more expensive than the 2MB cache xeons that are now available. SO u pay a whole lot more for the same performance at dual and less performance at quad. Also as u add processors opteron leaves the xeons behind due to the advanced bus architecture.

Re:xeons/opterons market share

[Loki_1929]

If you're interested in the architectual aspects, you really need to watch the presentation linked to by Ace's. It's an hour long, but it's well worth it if you're into that sort of thing. This is the 'father' of AMD64 talking about his baby, and then taking questions from reasonably intelligent people at the end.

Re:xeons/opterons market share

[Anonymous Coward]

What about a cheap XServe cluster with dual 64-bit G5s per node, 8GB ECC DDR RAM, dual 1GHz independent FSB per processor (instead of a single 400MHz for both like the Xeon), and they're pretty cheap.

Re:xeons/opterons market share

[cptgrudge]

but rather were upgrading to their high end dualie xeon systems -- they run very reliably, and very fast

I concur. I've got a dual Xeon 2.6 GHz (Dell) for use at work, and it's almost like the thing cannot be bogged down. I've tried. It just tools along blisteringly fast. It never goes down. I felt bad because I had to restart Windows to install a Java update (of all things) and lost my half year of Windows XP uptime. : (

Although, one of our Dell reps was out at WinHEC this year, and he said that those Opteron systems were really tight.

Re:xeons/opterons market share

[leandrod]

> I have not heard that the Itaniums were slow since Intel came out with the Itanium2.

Not slow, just irrelevant.

If one needs 64 bitness and high performance, there is RISC, already proven and more efficient. All free and (or) POSIX software runs there.

If one needs 64 bitness and x86 compatibility, there is AMD64, to run proprietary software it will take years to recompile and prove in 64 bits.

So...

A point that isn't made in the artical

[Pingular]

Xeon's are almost always for servers, wheras Opeteron's can be for anything. Try running a windows xp workstation on a dual Xeon system and you'll be very disappointed.

Re:A point that isn't made in the artical

[Anonymous Coward]

I call bullshit. You make a blanket statement without anything to support it or any logical argument at all. Of course you will get modded up, though, because your post is anti-Intel.

Re:A point that isn't made in the artical

[sheddd]

Xeon's are almost always for servers,

So are Opterons at the moment.

wheras Opeteron's can be for anything.

Both run x86 code pretty fast; Xeon can 'be for anything' too.

Try running a windows xp workstation on a dual Xeon system and you'll be very disappointed.

Only if I had to pay for it. Of course if I were in the market for a speedy dual workstation it'd most likely be an opteron due to price (and performance).

Re:A point that isn't made in the artical

[skiflyer]

I disagree. I'm running a Windows XP workstation with dual 2.4 GHz Xeons, and I'm not at all disappointed... neither are the 50 or 60 other developers surrounding me which are running on the same boxes.

What exactly would be our grounds for dissapointment?

Re:A point that isn't made in the artical

[hackstraw]

Xeon's are almost always for servers, wheras Opeteron's can be for anything. Try running a windows xp workstation on a dual Xeon system and you'll be very disappointed.

OK, lets go over this again. There is nothing really special about Xeons vs a P4 except the P4 is crippled so that it cannot do SMP, and there may be more cache options on a Xeon. Performancewise they are the same @ the same clock speed. FWIW, I've been dissapointed with XP regardless of the hardware :)

Now, back to this benchark thingy. 1st, I would appreciate in the article writeup that it said that it was only doing a simple read/write database benchmark, and that was it, but we don't come to slashdot for the stories, right? Also, in my opinion there was no significant difference between the two platforms regarding their speed on this benchmark. The difference between 1st and 2nd place, regardless of who won that test, was between 5 and 12%. I don't start to get interested until there is at least 20% difference, and even then that would only determine my choice for an initial purchase, I would never upgrade a system unless there was at least 100% speedup, preferably 200 -> 400% is worthy of doing an upgrade.

It would have been interesting to see results like this for more platforms, because I have not seen any significant numbers from the Opteron yet. For example, the memory bandwidth of the Opteron is 1/2 that of the Itanium2's.

Re:A point that isn't made in the artical

[Boba001]

Why would I be disappointed in my dual 2.4ghz Xeon workstation? There aren't many other choices for dual processor machines that don't cost an arm and a leg to build.

This machine totally destroys our dual P3 server or my single Athlon XP at home... Very happy with it.

Re:A point that isn't made in the artical

[TheRaven64]

Try running a windows xp workstation on a dual Xeon system and you'll be very disappointed.

I run Windows XP Pro (i.e. workstation) on a dual Xeon for controlling a VR haptic system, and I haven't been disappointed. Perhaps you could provide some evidence, or even reasoning, rather than just blanket statements?

Re:A point that isn't made in the artical

[theMerovingian]

I'm runnning dual xeon (2.66) on XP to do some pretty serious image processing stuff, and the performance is pretty good. All my bottlenecks are a result of crappy GIS software.

Re:A point that isn't made in the artical

[Pingular]

I disagree. I'm running a Windows XP workstation with dual 2.4 GHz Xeons, and I'm not at all disappointed... neither are the 50 or 60 other developers surrounding me which are running on the same boxes.
What exactly would be our grounds for dissapointment?
That your company spent $3750x2x55= $412500 on processors alone (assuming you have the 1mb MP model Xeons), when you could have the same performance for a quarter of that price.

Re:A point that isn't made in the artical

[Laur]

Also, in my opinion there was no significant difference between the two platforms regarding their speed on this benchmark. The difference between 1st and 2nd place, regardless of who won that test, was between 5 and 12%. I don't start to get interested until there is at least 20% difference

How about cost? The Xeons cost twice as much as the Opterons, and the Opterons give equivalent or better performance! Although you are correct that the performance difference may not be staggering (and between top of the line chips, who would expect it to be?), the price/performance ratio certainly is.

Re:A point that isn't made in the artical

[jefe7777]

you are right. both processors haul ass.

but the opteron, at all levels, will do it for a fraction of the price.

yes, that's right, not 5% less, not 10% less, BUT A MERE FRACTION...of the price.

just call 1-800-I-WANT-AMD

Call now and get your free Opteron beer chiller as a bonus gift....

(runs quickly away)

Re:A point that isn't made in the artical

[mobiux]

Why is it crappy?
We are looking at going to that where I work.
It's recommended by a consultant so we really don't have a lot of history with the product.

Any recommendations?

Re:A point that isn't made in the artical

[Greyfox]

I don't think I'd be disappointed. At least not once I formatted it over and installed Linux. Of course, I'd be upgrading from an Athlon 700...

Re:A point that isn't made in the artical

[cnkeller]

I don't start to get interested until there is at least 20% difference, and even then that would only determine my choice for an initial purchase, I would never upgrade a system unless there was at least 100% speedup, preferably 200 -> 400% is worthy of doing an upgrade.

Good post. However the one comment that I didn't agree with was the above.

My guess is that you aren't involved with any applications where compute time = money. When you are running simulations (say large CFD runs for example) that can takes days or weeks per run, a 50% improvement in speed is a major breakthrough if you get it by not touching code, ie hardware upgrades. Optimzing code is great and all, but it can introduce bugs and other expected behavior. Plus, us development people are pricey. Hardware is relatively cheap. Add in the fact that you generally get charged for CPU time on these big machines (or clusters of little ones), then *any* speed that you get is a major breakthrough, ie you can run more simulations in the same time for the same money.

In your environment, it's probably okay for you to only upgrade every three years when you get a doubling or more of performance, but there are enviroments where any speed increase is sought after highly, even if it's 20%. I suspect this is true of the special effects industry too, guys like Pixar, ILM, etc. If they can render more frames in the same time or even render the frames in the same time at a higher level of detail, that's worth paying for. Perhaps someone who knows more would care to enlighten us, I'm curious if I'm interpreting that correctly.

Re:A point that isn't made in the artical

[Anonymous Coward]

The Xeons cost twice as much as the Opterons

Even at pricewatch.com, that's blatently false. Furthermore, server customers care about the system price, not the CPU price -- Check out IBM.com (virtually the only Tier 1 AMD vendor):

x335: 2x Xeon 3.2Ghz, 1GB RAM -- $2,639.00
e325: 2x Opteron 248 2.2Ghz, 1GB RAM -- $3,399.00

Oh, lookie. the AMD system is actually more expensive. It's also a better product, so that makes sense. Retract your FUD.

Re:A point that isn't made in the artical

[TheLinuxSRC]

The 1Mb cache chips are very expensive, however regular 512Kb 2.4 Xeons can be had for under $200. The cache is very nice, but on a desktop workstation it is probably cost prohibitive not to mention, overkill.

Re:A point that isn't made in the artical

[Anonymous Coward]

A workstation would not have a Xeon MP model, trolling dipshit moron.

Nor would they have a Opteron 8-Series, which also cost several thousand bucks per CPU (more than the Xeon MP, in fact).

Re:A point that isn't made in the artical

[T3kno]

Any recommendations?

As a general rule: Get rid of the consultant :)

Re:A point that isn't made in the artical

[hackstraw]

My guess is that you aren't involved with any applications where compute time = money.

Your right. I work with scientists that run programs up to 5 days over 10 to 20 processors. We get our money upfront, but everyone wants their answers quickly.

When you are running simulations (say large CFD runs for example) that can takes days or weeks per run, a 50% improvement in speed is a major breakthrough if you get it by not touching code, ie hardware upgrades.

So your saying that its more cost effective for you to upgrade every 6 to 9 months? Thats fine if it pays off for you. You probably don't have that many processors to worry about either. Trust me, its not trivial to upgrade 60 to 120 processors that often, even if the machines were given to me.

In your environment, it's probably okay for you to only upgrade every three years when you get a doubling or more of performance, but there are enviroments where any speed increase is sought after highly, even if it's 20%.

Moore's "law" appears to still be holding true with a doubling every 18 months. Even I'm not slack enough to only upgrade every 3 years :)

Add in the fact that you generally get charged for CPU time on these big machines (or clusters of little ones), then *any* speed that you get is a major breakthrough, ie you can run more simulations in the same time for the same money.

CPU time charges are often proportional to the speed of the machine, and its cputime not walltime (there is a difference).

Re:A point that isn't made in the artical

[Deliveranc3]

Can we stop talking about the Xeon, until they come out with a new generation it's just garbage. If you think buying AMD risks your job then stick your head in the sand, the only thing that suffer's is your wallet and employer.

Let's talk about an issue that might still be relevant, so how about those new voodoo drivers!

Re:A point that isn't made in the artical

[dave420]

Errr...

$332x2x55=$36,520

Those are 2.8ghz Xeons, btw, so more expensive. They're also the price on their own, bought singularly. $37k for 55 developers CPUs is pretty good, and could undoubtedly be bought down to nearer $20k when bought bulk.

Seriously, price arguments about Intel don't work that well any more. They used to cost shedloads, now it's only their top-of-line processors (such as the extreme edition, and new models) that carry the weight (after all, if you need the fastest processor available, you can afford to pay for it :-P).

Re:A point that isn't made in the artical

[theMerovingian]

I guess I should quantify a little bit: It's the best there is on the market (I am actually a GIS consultant/ESRI vendor/etc).

The new 8.x implementation has been totally rewritten in VB, with the goals of looking pretty and being user friendly. The software runs pretty good, per se, but the performance is much slower than the previous GIS packages (command line ArcInfo 7 and ArcView 3.x).

Also, they have changed many of the menu interfaces, such that it is harder than it should be to migrate from the older software packages. For those of us 'old guard' GIS guys, who can do virtually anything using a combination of the command line interface and some AML scripts, it is much less productive to work in the new environment.

In short, if your staff doesn't have GIS experience, you will be perfectly happy using 8.x, and would be better off with it in the long run. And (mercifully) they are still including the command prompt functionality when you purchase an ArcInfo license.

If you only plan on using ArcView (the lesser license), you would probably actually prefer the new software.

An additional noteworthy feature of the new software (and by far the coolest aspect of it), is that ESRI finally figured out how to integrate ArcSDE into the core GIS package. SDE basically allows you to store all your GIS data in an enterprise database (Oracle and SQL Server primarily, I'm not sure about OS packages). The data load times are incredible, particularly for raster data (imagery). If you are a big company, or have numerous users, this is the way to go. Not cheap, but the performance is off the hook.

Let me know if you have specific questions, or if you don't like your current consultants :)

Re:A point that isn't made in the artical

[cnkeller]

o your saying that its more cost effective for you to upgrade every 6 to 9 months? Thats fine if it pays off for you. You probably don't have that many processors to worry about either. Trust me, its not trivial to upgrade 60 to 120 processors that often, even if the machines were given to me.

It depends, I'm not trying to make a blanket statement that this is always the case, but yes, I can certainly envision scenarios in that the benefit to customers is worth the price of the upgrade when you get less than a 200% return as you mentioned. I hope I didn't come across by saying uprade every 9 months for the latest and greatest.

As far as having too many processors, I work in one of the premier supercomputing centers in the world (shameless link) and we probably have more processors in a single computer than some people have at their entire sites. So I have *some* understanding of the logistics involved. Having said that, you're absolutely correct, replacing hundreds or thousands of CPU's isn't something that you do every year (or even every three).

As before, you make good points and I'm not really disagreeing with most of what your saying. Just trying to point out that it can make sense to upgrade when you get less than a 200% or 400% return on speed. Espeically when you have world-wide support dependencies, like providing the CPU time for the Return to Flight Initiative.

Re:A point that isn't made in the artical

[JudgeFurious]

I'd like to chime in here on why it's crappy. I don't use it myself but I support two GIS people here where I work. They're 2 out of 300 or so but I've noticed one thing.

No matter what they run it on it's dog slow. Those two guys get new machines, RAM upgrades, whatever they ask for, etc more often than any of my other users and I hear one thing consistantly from them. It's too friggin slow.

I don't know what the problem is. The guy here with the most understanding of it says it's the software and he's been working with it for years.

They need it, can't live without it, and love what they can do with it. They just hate it's performance.

Re:A point that isn't made in the artical

[ckaminski]

Is that with or without a warranty? :-)

Re:A point that isn't made in the artical

[mczak]

Dual systems usually use Xeon DP, which are a LOT cheaper than Xeon MP (not to mention faster due to the faster FSB, despite having less cache) (the same is also true for the Athlons, the 248 is much cheaper than the 848, in this case the chips though are identical safe the HT links).
That "Xeons cost twice as much" statement is correct for the Xeon MP 3Ghz/4MB cache vs. Opteron 848, which was the primary focus of this article.

Re:A point that isn't made in the artical

[radixvir]

The Xeons cost twice as much as the Opterons

??? I couldnt find the 4mb intels but the 1mb versions are alot cheaper than the opterons..

did i read your post wrong?

INTEL

AMD

Re:A point that isn't made in the artical

[Hoser McMoose]

I think that a large part of the story was simply that the new Xeons are no longer getting hugely beaten by the Opteron like they were in the previous test performed by the same site. While the Opterons still held a noticeable lead in 4P setups, the 2P systems were all pretty close.

In Anand's previous test comparing the two chips the Opteron came out WELL ahead of the Xeon. Of course, the tests were somewhat different, and it's been demonstrated several times that the Opteron is the chip to beat in web serving (subject of the first test) but things are much close in more pure-database tests (second article).

As for other platforms, I do with they had been able to throw an Itanium2 setup into the mix. In fact, what would be even more interesting is if/when they get Win2k3 64-bit edition running on the Opteron (and maybe Xeons as well) and THEN compare it to the Itanium2.

Re:A point that isn't made in the artical

[MrMr]

You won't be,
We're running four of those with Redhat 9 and fedora and with wildcat 6210 graphics as comp-chem workstations.
I don't think they're very slow at all...

Re:A point that isn't made in the artical

[imsabbel]

The 4MB Xeons MP are MUCH more expensive than the normal dual Xeons.
One 3Ghz XEON MP cost about 3500$

Re:A point that isn't made in the artical

[Anonymous Coward]

FWIW, I've been dissapointed with XP regardless of the hardware :)

XP: able to cripple the fastest processor ;)

Also, in my opinion there was no significant difference between the two platforms regarding their speed on this benchmark.

Man, do I disagree! Yeah, the performance of the systems was close, as long as you ignore the fact that the AMD's were running at 2/3 the clock speed with 1/4 the cache! AMD clearly has a better design than Intel. All of the technology that Intel has will eventually belong to AMD, also, and their design obviously takes better advantage of any technology!

Has anyone had a look at the datasheets for AMD and Intel? Does the lower clock speed translate into an equivalent savings in power conmsumption? Or is that offset by the larger geometry?

Re:A point that isn't made in the artical

[fitten]

Man, do I disagree! Yeah, the performance of the systems was close, as long as you ignore the fact that the AMD's were running at 2/3 the clock speed with 1/4 the cache!

Who cares. You cannot compare two different architectures on a clock per clock basis. It makes no sense.

AMD clearly has a better design than Intel.

Purely subjective.

All of the technology that Intel has will eventually belong to AMD, also, and their design obviously takes better advantage of any technology!

This statement makes no sense at all. AMD does not make Borg chips.

I would imagine that you actually know very little about actual CPU inner workings but obviously have loud opinions on which you have no clue.

Re:A point that isn't made in the artical

[dcam]

Not only that, but when they were comparing prices, they said that the 2Mb cache Xeons were twice the price of the Opterons. Howevere in the article they tested the *4Mb cache Xeons*, which you can expect to be even more expensive.

In terms of bang for your buck Opterons rock.

Re:A point that isn't made in the artical

[theMerovingian]

They need it, can't live without it, and love what they can do with it. They just hate it's performance.

That about sums it up! It is so useful, though, that the business value overrides the poor performance in most instances.

That being said (and to get this back on topic), my dual Xeon workstation is a pretty darn good GIS machine. Mine has a gig of ram and a 250gig scsi drive, which is more than adequate for most high end use. It would be nice to have an extra stick of ram for some of the big datasets.

Also, a dual monitor video card does wonders for GIS, because you can have your map on one screen and your tables on the other.

Re:A point that isn't made in the artical

[Bloater]

Unfortunately I couldn't see the graphs. I think the important measurements would be queries per second per dollar, energy consumed during benchmark, and price over lifetime.

Re:A point that isn't made in the artical

[Noren]

Okay, I checked out pricewatch.com, and you're wrong about the 8 processor machines (requiring Xeon MP rather than Xeon DP processors) which the article and grandparent post talked about.

No Xeon MP as fast as 3.0 GHz or with the 4MB cache is available via pricewatch.
The fastest Xeon MP processor available they list is 2.8 GHz with 2MB cache, for $3788
A single Opteron 848 processor costs $1469

Oh, lookie, the AMD processor is less than half the price of a Xeon product inferior to the one that article benchmarked it against. Retract your lies.

Re:A point that isn't made in the artical

[skiflyer]

That would be a pretty odd assumption since I mentioned they were developer workstations.

The current incarnation of the whole machine goes for about 1800$ (with 3.2GhZ chips), I'm not so dissapointed in that. I'm not saying it would be where I'd put my pennies if I were buying a machine at home, but for a corporation who needs stability and uniformity in an MS Windows environment, I just don't see a clear "they made the wrong choice" argument.

Re:A point that isn't made in the artical

[Anonymous Coward]

they are if you factor in price (as you should)

build a cluster, and your on a budget

4mb cache zeons are $4000 each. and those are the only ones that really hold firm across the board against opterons.

the 2mb zeons are still TWICE the cost of the fastest opterons.

i don't understand wtf is so hard. this is simple math.

all the performance, half the costs.

if you spend a million dollars on cpus for a zeon cluster, or you can spend $500,000 on cpus for an opteron cluster.

Re:A point that isn't made in the artical

[hyc]

Systems that charge for CPU time do it based on number of machine/instruction cycles. If you have two machines of identical architecture, one running at a faster clock rate, then both should solve an identical problem in an identical number of cycles. In these cases, the machine provider will generally charge a higher price per cycle for the faster machine.

In today's competitive world "time = money" is always true, but some kinds of time are more expensive than others. Buying time on the fastest processor isn't always the right choice.

Re:A point that isn't made in the artical

[cptgrudge]

Try running a windows xp workstation on a dual Xeon system and you'll be very disappointed.

Untrue.

I'm using a dual Xeon system right now with Windows XP Pro as my OS. It fucking screams. I'm not disappointed in the slightest with it.

Re:A point that isn't made in the artical

[Anonymous Coward]

Who cares. You cannot compare two different architectures on a clock per clock basis. It makes no sense.
You can't? Then just what do you compare them on? This is a very real measure of just how much work is being done per clock cycle. There are physical limits to how fast you can clock any hardware, so the more efficient any processor is during each clock cycle, the more speed you can get out of the processor when you can't carnk up the clock rate anymore.

Purely subjective.
Purely objective; AMD is doing more work with less hardware resources. What the hell constitutes a better design to you?

This statement makes no sense at all. AMD does not make Borg chips.
Obviously a trekkie, you can't think outside of stupid movie references. Intel, right now, has better fab facilities than AMD does, which lets them build on smaller geometries and run at higher clock rates. AMD will eventually build new fab facilties to stay competitive, but it takes time. My point in all the above comments was that, when that happens, AMD's offerings will run about 1.5 times faster when both designs top out at their fastest clock rates.

I would imagine that you actually know very little about actual CPU inner workings but obviously have loud opinions on which you have no clue.
Pot, kettle, black!

Re:A point that isn't made in the artical

[Knetzar]

Intel is able to get the P4s to run at a higher clock rate then Opterons because the P4s have a longer pipeline. Even with equivalent manufacturing techniques P4s will run at higher clock rates then Opterons.

Quick Link to Test Results

[hng_rval]

Test Results can be found here:
http://www.anandtech.com/IT/showdoc.html?i=1982&p= 6

Re:Quick Link to Test Results

[Anonymous Coward]

"An area that we didn't touch on is cost, which is where AMD truly shines. The Opteron 848 processors we tested are around 1/2 the price of Intel's 2MB L3 Xeon MPs and we have not seen retail data on how expensive the 4MB parts will be."

In other words, you should be able to buy a quad-Opteron set up for roughly what a dual Xeon 2MB set up. The difference will be even greater with the 4MB chips.

Quad-Opteron being their favorite, I have to say, go AMD!

summer cottage?

[SuperBanana]

for-your-summer-cottage's-data-farm

Ah, so for all our college-student friends, that would be "the parents' house"?

Re:summer cottage?

[Monkelectric]

If only there was a "+1 I feel Shame"

Cache always help

[Anonymous Coward]

I remember AMD's K6-3 would blow away the K6-2 at the same clock speed with the major difference being the cache.

Re:Cache always help

[NeoTheOne]

Yeah when you're comparing a 4mb cache processor against a 1mb cache processor what do you expect...not exactly rocket surgery people ;-)

I want a 4mb opteron tho :D

Re:Cache always help

[redshadow01]

Dude, read the article. The new Xeon has a 4meg L3 cache, but its effects are limited because the FSB is only 400Mhz.

Re:Cache always help

[Anonymous Coward]

And? The article says they use the cache because of the FSB bottleneck, which means the CACHE HELPS performance. That was my point. More cache = better performance.

Re:Cache always help

[Vaystrem]

Cache may always help but this is not as straightforward a statement as you indicate. It is highly dependent upon the architecture of the processor.

The reason the 4mb Xeon's are significantly outperforming the 2mb Xeon's is due to the shared bandwidth architecture of the Xeon's. The cache makes up for the lack of access to data via the FSB and keeps the very deep pipeline of the P4 series processors full. The long pipeline is the reason that cache misses impact the speed of the P4s so much - despite Intels attempt to improve branch prediction. Simply look @ the P4 Celeron's to see how they can be so utterly trounced by regular P4s @ the same clockspeed with little architectural difference but cache size.

Opterons/AMD 64s do not benefit as much from the boost in L2 Cache. Perfect example of this is the
Athlon 64 3000+ and Athlon 64 3200+

The 3200 has 1meg of L2 - and the 3000 has 512k - and both run @ 2 ghz. The performance difference between these two (in most benchmarks) is less than 10%
Anand Review of Athlon 64 3400+

So a doubling of cache at the same processor speed results in a 10% boost in performance 'maybe'.

Finally some applications are more sensitive to L2 cache sizes then others.

Therefore your statement "more L2 cache always helps" is strictly true - but the degree of performance increase must be compared against the increase in cost. And this benefit will change from processor to processor and application to application.

Re:Cache always help

[Anonymous Coward]

K6-2 to K6-3 was dramatic. Thoroughbred to Barton was not. Athlon64 3000 to Athlon64 3200 is even less dramatic (maybe 1-3%?).

Cache does not always help. Once your critical path fits in cache what more can you do? The very interesting thing is that cutting the K8's L2 in half (3200->3000) results in only a couple % performance degradation.

But scaling the K8's MHz up 10% (3200->3400) results in 10% speedup in benchmarks--almost exactly.

This, along with the upcoming trend in multi-die CPUs and hyperthreading, points that we are not as bandwidth-limited as we've been told anymore. And we can't push MHz faster yet. So for now we need more parallelism in the CPU itself.

Clearly right now MHz and ALUs are what really count for speed.

I recommend Glasses

[Avrice]

Whomever is citing Anandtech as claiming the dual Xeons almost always beat the dual-Opterons needs to read the article again. Both Architectures in a dual configuration tended to perfom about the same with Opteron and Xeon each winning some of the time. The Opteron scales better above dual configurations. However the Opteron is HALF the price of a Xeon! Cost/performance (or else we would all have 12th generation DECAlphas or Power5s by now) is easily handed to Opteron. Nice spin!

Re:I recommend Glasses

[Anonymous Coward]

You're absollutly right. The few results that Xeon won was by a little margin and were allways in 2-way configuration. In 4-way Opteron cannot be beated... An remember top Opteron is 2.2GHz while Xeon is 3GHz... Damn, my windows machine (I have to use frontpage :-( ) has 800MHz!!! Besides being a cost/performance suberb ratio it's an even better performance/MHZ... "It's not the size that methers, it's what you do with it" (in this case, clock speed ;-) )

PS - Sorry for any bad speling, but I'm sure I write better english than you write portuguese :-P

Re:I recommend Glasses

[Hektor_Troy]

I suggest using Opterion contact lenses like this poster.

Re:I recommend Glasses

[embarcadero]

In addition, Anand used sub-optimal memory in the Opteron, and non-NUMA config. Looks like he had some Intel "assistance" in designing the "benchmarks" as well... the database read/write ratio is not at all realistic, favors the Xeon.

Re:I recommend Glasses

[ms8423]

Thanks for pointing it out. From the conclusion of the article:

The comparison we've made here is a very important one; it identifies Intel's strengths and their weaknesses with Xeon, and it crowns Opteron a clear multiprocessor winner. An area that we didn't touch on is cost, which is where AMD truly shines. The Opteron 848 processors we tested are around 1/2 the price of Intel's 2MB L3 Xeon MPs and we have not seen retail data on how expensive the 4MB parts will be.

In a 4-way configuration AMD's Opteron cannot be beat, and thus it is our choice for the basis for our new Forums database server. We'll be documenting that upgrade in a separate article so stay tuned.

Not quite as the /. story reads.

Re:I recommend Glasses

[Anonymous Coward]

However the Opteron is HALF the price of a Xeon!

Heh, you haven't been to Pricewatch lately. An Opteron 248 is $949 compared with $821 for a 3.2ghz Xeon (1mb cache). Dell has the 2mb Xeon listed at $1,499. We know the street price will be a couple hundred lower.

AMD can't afford to sell at HALF the price of Intel.

Re:I recommend Glasses

[Anonymous Coward]

No surprise there. Very few of these popular "review" sites have anyone on staff (or even freelance) who know anything about computer architecture. These guys are all hobbiests with backgrounds based in futzing around with their PC and reading press releases. They have zero understanding, or "deep knowledge" of the design trade-offs and implementation details that make all the difference in these systems.

With few exceptions these guys are only good for putting pretty lights in their computers and running benchmarks. But in no way are they qualified to CHOOSE appropriate benchmarks nor to evaluate the results and draw conclusions beyond the raw numbers.

For the most part, it isn't their fault, it is a classic case of, "Those who can do, those who can't teach." The people with the ability to properly evaluate systems can make a ton more money working in the industry than they could writing articles for free websites.

Re:I recommend Glasses

[Anonymous Coward]

First of all, did you read the Anandtech article instead of looking at just the graphs. If you did, you would realize that a Xeon with 1MB L3 would get thrashed silly. Xeons absolutely need as big L2/L3 as possible to overcome the shared bus bottleneck. Bringing up the price of a "crippled" Xeon w/ 1MB means nothing.

Here's the cheapest prices I could find on Pricewatch for 2MB Xeons:

Xeon 2.0GHz: $1269
Xeon 2.8GHz: $3788

Sure looks like street prices will be a couple of hundred lower...NOT!

Re:I recommend Glasses

[Gat1024]

If you read the article you would see that, where listed, the Xeon 3.2 ghz 2mb is the speed king. It's performance is virtually identical with the 4mb (slower bus) part -- and cheaper to boot. Anandtech says as much in their conclusion.

Go to Dell and they quote $1,499 for a Xeon 3.2 2mb as a second processor in their servers. Everyone knows that you don't buy your second processor from Dell because you pay a premium. What you're probably getting from Pricewatch is a bunch of hits from someone's pre-pricedrop processors. You will have to wait until the new prices get into the channel.

Re:I recommend Glasses

[Anonymous Coward]

Item #1, The 2MB shown at Anandtech are the new 533MHz FSB Xeons. They will have an extra premium over the current crippled 400MHz FSB ones.

Item #2, the pricewatch figures I listed were for old Xeons. They were introduced a year ago? Are you still waiting for the "street" prices to appear for them? I have a suspicion there's no such thing as street price discounts for high-end Xeons.

Re:I recommend Glasses

[Jeff DeMaagd]

Looks like he had some Intel "assistance"

Which is pretty interesting as I've read several slashdotters claim that he's pro-AMD.

Re:I recommend Glasses

[brucmack]

RTFA. Anandtech did not "choose" a benchmark for this test, they developed two themselves based on real applications. The average-load test ran queries on their forum system, and the high-load test ran an enterprise system in use by an anonymous company.

As for them not knowing anything about computer architecture, you should read some of the articles that have been put up on Anandtech when new CPUs and GPUs have come out. They go into very good detail about exactly how the chips are working, and as a computer engineering major I can say that very little of it is flawed information.

I know that some review sites are as you describe, but I do not believe Anandtech to be among that group. YMMV.

Re:Why benchmark games?

[Lomby]

Hmmm, you should read the article before commenting.

The last two articles on Xeons used their forum database as the workload for the benchmark. In the current article he even managed to use an unnamed enterprise order management system.

Then, if you have the games and the 64 bits systems at hand, why not do a quick benchmark?

Their review of windows64 highlighted some obvious problems, probably with drivers/PCI, that may be relevant for professional use (think of CAD).

NUMA means Opteron is Better

[Anonymous Coward]

At two processors Xeon is still ok because the bandwidth of the memory coherency still isn't in serious contention. However, as the systems scale larger support for NUMA is critical to reducing memory latency because it means that memory does not have to flow in from the controllers on other processors.

That is why Opteron is required for good performance with eight to sixteen processors, and you can even see the improvement on the four way tests that Anand ran.

Re:NUMA means Opteron is Better

[Anonymous Coward]

NUMA is rarely employed in systems at the 8 to 16 processor size. Those are typically SMP.

Saying Opteron is better for 16 way means nothing as those systems do not exist.

NUMA architectures are not without issue. I emphasize the "NON UNIFORM" aspect of the acronym. Even if SGI wants to change it to "nearly uniform". Sounds like you've been reading too much of their properganda.

Re:NUMA means Opteron is Better

[Anonymous Coward]

marijuana does not kill brain cells or cause permanent brain damage. alcohol does. long term consumption of alcohol has shown consistent loss of brain mass (convoluted surface space). a joint is less harsh and harmful than a cigarette. all of the harmful properties of smoke inhalation can be avoided by using a vaporizer to inhale THC. THC only causes a lower sperm count while under the influence, but never after. everything the government shoves down your throat that it calls facts are lies. i hate it when it releases some dumb survery that says "ecstacy" is losing popularity. maybe, but that does reflect an accurate picture. more popular drugs are entering the market such as 2c-i. it does not reflect the drugs that they bash all day long by lies and misconception. the government is lying to you. fight thelies!

Re:NUMA means Opteron is Better

[flaming-opus]

You will find that most high-end Xeon systems are also NUMA systems. IBM, Unisys, HP all construct their really big xeon boxes as NUMA-clusters of 4-processor SMPs. They create a distributed memory machine at the chip-set level. This is actually what the opteron does, except that the chip-set (well, the memory controller part of it) is built into the processor.

I think the above poster had the correct idea about NUMA, but worded it in a misleading way. A NUMA design (either of opterons, or of Xeon-quads) will have to do some memory access through the memory controllers on other nodes. This increases the latency of memory access, and can clog up the inter-processor links if lots of memory loads/stores go to remote memory. Thus NUMA-aware operating systems and system libraries are necessary to maximize the amount of memory access that is local, and minimize the usage of the inter-processor links.

While the opteron design is elegant, and fast, it is not the only smart way to do things. It offers great aggregate memory bandwidth, but can slow things down in the worst case. Most large NUMA systems are created by linking 4-way SMP nodes. (Examples: Sunfire, HP alphaservers, Cray X1, NEC SX-6, Unisys 7000, IBM xseries 4xx xeon, IBM xseries 4xx itanium,...) Apart from opteron systems, the only systems I can think of that do NUMA per processor are the cray T3E, SGI origin, and intel paragon, all of which are Massively parallel supercomuters.

It is safe to say, however, that a shared bus system does not scale well beyond a few processors. This is best demonstrated by the 36 processor SGI-challengeXL, which was significantly bottle-necked at the memory bus.

food for thought.

Re:NUMA means Opteron is Better

[Jeffrey Baker]

Every n-way Opteron where n > 1 in the field is a NUMA, so your "rarely" no longer applies.

Re:NUMA means Opteron is Better

[Mr. Piddle]

Actually, I'm pretty sure all big SMP systems are NUMA, in a sense. Opteron is similar to UltraSPARC in that each CPU gets its own local RAM, and the CPUs are networked together in some manner with a rediculous-speed interconnect (GB/sec rather than Mb/sec for Ethernet). I belive such a network might even be called HyperTransport by AMD.

Re:NUMA means Opteron is Better

[Fnord]

I don't think you really understand how NUMA works. The whole point of NUMA is that memory DOES have to flow from controllers on other processors. Xeons use a uniform memory architecture, meaning that they all share a memory controller. Even if that memory controller is faster than a single proc system's memory controller, they're still contending for it. However, they do all have equal access to it.

NUMA is a tradeoff. Each processor has its own memory controller and its own bank of memory. Therefore each processor has preferential access to that bank of memory. If, however, a processor attempts to access memory in another processor's bank, its slower. This cost is mitigated by intelligent VMs that attempt to put the memory for processes in the memory bank controlled by the processor running that process. If this is done efficiently enough, the benifit of having 4 memory controllers far outweighs the cost of possibly having to get memory slowly from another processor.

Re:NUMA means Opteron is Better

[Anonymous Coward]

URL to a sixteen-way Opteron box, please?

Oh that's right, they don't exist. :*

Re:NUMA means Opteron is Better

[Anonymous Coward]

Cray is using 1000's of Opterons in their supercomputers. But yeah, they use connecting hardware.

Re:Why benchmark games?

[Anonymous Coward]

How could this get modded up? The article doesn't use game benchmarks at all. The poster obviously just wanted to get the first meaningful post and didn't even bother looking at the article.

Re:Why benchmark games?

[Judg3]

It's probably due to the lack of knowledge/tools to benchtest anything else. I'd like to see SQL benchtests, IIS/Apache test/etc but just like a lot of other people, I don't know exactly how to do that. Though if I ran a site which made it my business to test hardware I'd definately find out and learn how to do it.

I'd like to see more "Consumer Reports" type tests to. Test hardware configuration X as a high-volume SQL server, and show me how it's held up after a month, 3 months, 6 months, and a year. Yes, maybe I'd upgrade before then, but not everyone would, and I'd like to see common failures and problems down the line - not a 1-2 day test.

Can someone please clear my ideas about this

[SlashingComments]

I saw this yeasterday on his site. Pretty good.

One thing I did not understand is how come the 3MB cache is helping with big database query ? I thought that will thrash the cache and there will be not much performance gain if you are working with bigger code/data set. Also, for the four CPU opteron, do they have hyper transport going from every cpu to every cpu ? Is it like a mesh or like a ring where every cpu has only two connections to it's next ones.

Another thing I did not get is how linux is handling ( not handling ) the local memory to the CPU. This thing looks like a mini-numa type system. Does linux actually try to keep the data in the RAM and process it with the cpu it is connected to ? how does this really work ?

May be you guys can help clear my ideas .

Re:Can someone please clear my ideas about this

[chef_raekwon]

ne thing I did not understand is how come the 3MB cache is helping with big database query

this is interesting, and i don't have an answer, except to say that SQL servers generally try to load all of the tables into available RAM. If the data is too large, then simply the indexes(??). If the server ever has to go back to the harddrive for data (which would make it bloody slow for a query) it will check recently cached stuff first - and larger caches means reduced time pulling data sets from a raid array or single harddrive.

that is atleast my take...it generally differs from server to server....MySQL does not run exactly the same way as MSSQL as Oracle. (which means I've generalized.)

Re:Can someone please clear my ideas about this

[SunBug]

L1,L2,L3 caches relate to caches either on the CPU or very close to it. Only when a bit of code isn't found in the L1,L2[,L3] cache, does the processer go all the way out to the system memory to fetch the next set of instructions.

Now, a database generally runs a fixed set of instructions during a query, though those instructions may not fit in the smaller L1 and L2 caches (the L2 on the dual Xeon we just bought was 512k, but 1mb is also available). Having the larger L3 cache can help keep those instructions for doing queries closer to the CPU, which can significantly speed things up.

As for database design, good database will use available index(es) when possible during a query since they shortcut looking for data. This is why having proper indices is important to maximize database performance.

The database will use as much memory as you allow it to cache data. Only when the data isn't resident in memory does it go out to the HDDs to find it, regardless of whether the query was able to make use of index(es).

Both databases and CPU caches make use of MRU/LRU (Most/Least Recently Used) to determine what to keep and what to throw away. The MRU stuff stays, and the LRU stuff is flushed out when something new is stored in the cache/database memory.

A complex SELECT statement that returns millions of rows, but makes use of 2mb of instructions would significantly improve performance with the addition of a 3mb L3 cache since the CPU wouldn't have to go to system memory to find the next set of instructions, per row. Save a few ms here and there fetching instructions, and it adds up over time.

Re:Can someone please clear my ideas about this

[flaming-opus]

They don't have hypertransport from every CPU to every other. An opteron has 3 ht links. Since some of those need to be used to connect to the system I/O devices, you have 2 left for your mini-numa system. Thus the processors would have to be connected in a ring.

Linux is capable of intelligent memory layout. It can migrate data to the processor on which the threads are running, is intelligent about which processor runs which threads, and can make duplicate copies of read-only data. It works reasonably well. (some of this is the stuff SCO is in a huff about) However, I doubt this functionality is turned on in any off-the-shelf distros. If the benchmarker compiled a kernel with NUMA in mind, this would work, otherwise I doubt it.

Incidently, since the two streams of a hyperthreading-capable xeon share the L2 and L3 caches, they benefit from NUMA grouping also.

Re:Can someone please clear my ideas about this

[Anonymous Coward]

One thing I did not understand is how come the 3MB cache is helping with big database query ? I thought that will thrash the cache and there will be not much performance gain if you are working with bigger code/data set.

You are correct.

I recently wrote some matrix code in C. Play with the sizes of drows, dcols, mrows, and mcols some.

/*ht.c
*This is a C program to test cache hits/misses
*using matrix multiplication.
*/

#include <math.h>
#include <stdlib.h>

#define DROWS 256
#define DCOLS 128
#define MROWS DCOLS
#define MCOLS 256

int
main(void)
{
unsigned int seed = 8123;

int i,j,k;
double d[DROWS][DCOLS];
double m[MROWS][MCOLS];
double r[DROWS][MCOLS] = {0};

for(i = 0; i < DROWS; i++)
for(j = 0; j < DCOLS; j++)
d[i][j] = rand_r(&seed) / RAND_MAX;

for(i = 0; i < MROWS; i++)
for(j = 0; j < MCOLS; j++)
m[i][j] = rand_r(&seed) / RAND_MAX;

for(i = 0; i < DROWS; i++)
for(j = 0; j < MCOLS; j++)
for(k = 0; k < MROWS; k++)
r[i][j] += d[i][k] * m[k][j];

/*
for(j = 0; j < MCOLS; j++)
for(i = 0; i < DROWS; i++)
for(k = 0; k < MROWS; k++)
r[i][j] += d[i][k] * m[k][j];
*/

return EXIT_SUCCESS;
}

Compile and run it using time under bash:

bash-2.03$ gcc -o ht ht.c
bash-2.03$ time { ./ht; }

real 0m2.108s
user 0m2.080s
sys 0m0.010s

.

Vary the sizes of the matrices and swap to the second (commented out) multiplication routine. They produce the same result, but one will generate more cache misses than the other (if you get the matrix sizes right) resulting in a 50% slowdown on an Athlon XP 1900+ or an unintelligible crap pile on a Sparc running Solaris version unknown.

Make the matrices too small and you can't tell the effects of the cache because the operations are too fast. Make the matrices too large (not hard on a Sparc, as I just learned) and the cache hits/misses won't make a noticeable difference in time, assuming you're not using a Sparc and you succeed in not segfaulting.

Re:Can someone please clear my ideas about this

[Twillerror]

SQL servers use a page caching system generally. That is the database exists on the harddrive as a series of xK pages ( 8k pages for MS SQL server ).

As a page gets loaded from the harddisk it is loaded into the server's cache. Any read/writes are done to memory and not the disk. Background process write the pages to disk that are dirty. All transactions are written to the transaction log so if the server crashes before this happens recovery can happen when the db starts back up.

This means that a large portion of the data is already in memory. Servers usually pre-allocate gigs of memory for this purpose, the more the better and a big reason for 64 bit on large dbs.

Under x86 caching schemes, the CPU does speculative loads. It "guesses" what memory the processor is going to need and starts loading it into high speed cache. This is perfect for a db since most of the time the db pages you need are in sequential order. Especially when you are talking about pages that only include indexing data. The query usually does most of it's work using indexes, and then at the end will actually "lookup" the data.

So bigger caches mean that these big binary trees get loaded into cache and the algorythms can loop through them faster and pull off the cache.

Take this into the Itanium world and we can start to get even better performance. The thing people tend to forget about Itanium is that you can tell it to load you data into cache. So an optimized DB server can have it read this large section of code into the cache while it does calculations. Itanium allows 3 instructions to be loaded. Once hypertheading is put into Itanium you will see these DB apps really fly. Itanium is showing good promise in this arena, even at 1.5 GHZ. Clock that up to 2 to 3ghz with multiple hypertheaded cores and we are going to have one fast chip for dbs.

The big issue is the price, if you are going to spend that much money, go with the proven Sun/IBM. Itanium is set to replace the Xeon by 2007 ( I'm guessing before then because of scaling issue on the Xeon and x86 emulation software giving decent 32 bit for legacy apps ).

I really think Intel needs to push their Itanium. MS likes it, Linux likes it, a few db servers like it, and a slew of other high performance, server based things. I don't see how Intel is going to scale against the Oppie. Not unless they stick a on-die memory controller. Hypertheading and the new thread based instructions will help though. Should be an inresting battle. I'll be happy if AMD can get 10% to 20% marketshare, then we will see some true competition and innovation like we have on the desktop.

Re:Why benchmark games?

[Anonymous Coward]

Did you even read the review? Why don't you read before you end up making yourself look bad.

And Anand's Win XP 64 review wasn't bashing it. He was using games to highlight that there are still several things that need to be addressed (Like drivers) before it will be ready for prime time.

Importance of compilers

[Anonymous Coward]

Believe it or not, Intel's compiler generates very good code for the Opteron. Far better than GCC or generic IA32 compilers.

So in any evaluation, the compiler and binaries that are used is an important question.

There was no mention of this in the article.

Re:Importance of compilers

[Anonymous Coward]

i use skyos's compiler. much better performance.

Re:Importance of compilers

[Aardpig]

Believe it or not, Intel's compiler generates very good code for the Opteron. Far better than GCC or generic IA32 compilers.

This is the experience I've had with the Intel Fortran compiler (ifort) on an Athlon XP. Codes compiled with ifort are around twice as fast as those compiled with GNU g77 (for Fortran 77), and around 1.5 times faster than those compiled with Lahey lf95 (for Fortran 95).

Re:Importance of compilers

[Hoser McMoose]

The difference isn't all that large actually, though ICC is generally still a bit faster than GCC on Athlon64/Opteron chips. If you look at the SPEC CPU2000 results you can find a bunch of different Opteron results. Some of these results can be used to show a direct comparison between ICC and GCC, and they're often within about 10% of one another. In fact, when you start dealing with 64-bit Linux and 64-bit GCC code it often ends up being faster than 32-bit ICC code.

As for the compiler used in the article, they were using Microsoft SQL Server as best as I can tell, so I think that you can be fairly certain that the compiler used was MS Visual Studio.Net.

OS

[millahtime]

So I see that M$ Windows was used as the OS. Unless this was a prerelease of the 64bit XP then they were running a 32bit OS on the chips. So, wouldn't that mean that this isn't a true test of the power?? Your not taking full advantage of the 64bit power.

Re:OS

[DangerSteel]

Specifically the article says Windows 2003 Enterprise server of which there are at least two 64 bit versions, but they don't tell you which version they used. I don't always agree with their findings but they appear to be sharp people for the most part. I doubt they would ever use XP for a server class chip test.

Re:OS

[flaming-opus]

Most opteron users are also going to be running 32-bit OSes, so benchmarking them as such is probably more usefull than otherwise. Additionally it's nice to see benchmarks that attempt to test with 1 variable (hardware) instead of comparing microsofts variations of software as well.

The fact that opteron is 64bit-capable is cool, but in performance analysis it's the underlying architecture that matters, not the register width.

Back to Intel Fanboy

[superpulpsicle]

Alright I have had about 3 AMD processors die on me. I have owned about 4 Intel processors all the way back from original Pentium. Not one has ever had a problem.

Now... given this kind of statistics, as sad as it may sound I'd say I am willing to pay anything for an Intel just to avoid the headaches.

Re:Back to Intel Fanboy

[DAldredge]

WTF do you do to them to make them die?

Re:Back to Intel Fanboy

[tuffy]

I've owned 5 AMD processors from the K5 to an Athlon64 and all are still in perfect working order. But these sorts of anecdotes aren't very helpful in determining average chip reliability.

Re:Back to Intel Fanboy

[hng_rval]

Alright I have had about 3 AMD processors die on me. I have owned about 4 Intel processors all the way back from original Pentium. Not one has ever had a problem.
Now... given this kind of statistics, as sad as it may sound I'd say I am willing to pay anything for an Intel just to avoid the headaches.

That is an interesting use of the word statistics. In order to determine if your next processor is likely to break, you should look at thousands or hundereds of thousands of Intel procs and AMD procs. Your 7 processor study is inherently flawed.

Re:Back to Intel Fanboy

[hyperstation]

you know, you're supposed to outfit those AMD processors with fans, heatsinks, and some of that thermal paste....

Re:Back to Intel Fanboy

[John Courtland]

Wow, I have a working AMD 386/40. Yet I have a score of dead Intel 286/386/486's. I just evened out your "statistics". Not to mention the 5th gen and above x86 class processors I have.

Re:Back to Intel Fanboy

[sp0rk173]

Fascinating case of perception vs. reality. I have four AMD systems that are rock solid and have been for atleast a year (in some cases, one processor goes back atleast 5 years and it's still running strong to this day as my router). I've also worked on my cousin's computers, which are all Intel-based, and until I bitched him out, he was calling me every other week with problems he was having - yet he stood by the mantra of "Intel is more stable, and worth the extra money." Given these "statistics," providing a mirror image of yours, it seems intel isn't worth the extra money. Of course, these statistics aren't statistics at all, rather fallbacks to consumer dogma and vendor-conditioning. Either way, do what makes you feel comfortable in your choice, and keeps you consuming. It's the American Way.

Re:Back to Intel Fanboy

[Mr Guy]

And, as we all know, the plural of anecdote is indeed "data".

Re:Back to Intel Fanboy

[irokitt]

I don't really buy into your reasoning. 7 processors does not a survey make.
I know my Dad refuses to buy anything other than Intel silicon because of his past experience-buying a Cyrix 6x86 and getting burned because of it.
Other than the fact that the Athlon XP series did not have a heat spreader and could be crushed when applying a fan too clumsily, I have not heard any reason for an AMD chip to fail faster. Assuming, of course, that you take the normal precautions-heatsink w/ fan and grease, a good cooling system in your case, and no botched overclocking attempts.

Re:Back to Intel Fanboy

[Anonymous Coward]

He/she may have bought them from the same manufacturer all about the same time (which would make sense e.g. replacing failed hardware). So, a specific batch could have sucked. I remember something about capacitors on MBs being bad so why not a chip?

His "statistics" probably wouldn't make it to the peer review process. *chuckle*

Re:Back to Intel Fanboy

[Anonymous Coward]

> I have not heard any reason for an AMD chip to fail faster

Well, AMD has historically marketed itself as a "budget" processor. When put in a "budget" system, it's going to be mated with cheap mobos and cheap RAM and cheap power supplies, all of which have higher failure rates.

Unfortunately AMD gets a bad reputation from this, when it's probably not their fault.

Re:Back to Intel Fanboy

[Anonymous Coward]

Odds are the problem has nothing to do with the cpu. Occams razor suggests superpulpsicle is in fact the problem.

Re:Back to Intel Fanboy

[crabpeople]

maybe you need more fans

fanboy.

Re:Back to Intel Fanboy

[lukateake]

Alright I have had about 3 AMD processors die on me. I have owned about 4 Intel processors all the way back from original Pentium. Not one has ever had a problem. Now... given this kind of statistics, as sad as it may sound I'd say I am willing to pay anything for an Intel just to avoid the headaches.

That is an interesting use of the word statistics. In order to determine if your next processor is likely to break, you should look at thousands or hundereds of thousands of Intel procs and AMD procs. Your 7 processor study is inherently flawed.

Actually, one can start to get a statistically significant result at around as little as 30 data points depending on a few factors.

For giggles, let's take a large population (the MS Windows install base: ~300M) and if I wanted to be 95 percent (SixSigma!) sure that I calculated within a confidence interval of plus/minus 5 points, how many must I sample? Thousands? Hundreds of thousands?

Nope, a scant 384! A generally accepted confidence interval is +/-4 at 95% and you see this used in several polls. To get to that level, I'd need to sample only 600 of the 300,000,000 population.

See for yourself.

Re:Back to Intel Fanboy

[Monkelectric]

une question pour vous -- wtf are you DOING to them? I've only seen one system of ANY kind die of cpu failure in 15 years of working with computers -- and that was a Sparcstation 5 that hadn't been turned off in 10 years.

If you're telling me you've had 3 AMD's die that weren't overclocked, weren't modified, and had proper cooling with good power supplies, i would be very surprised.

Re:Back to Intel Fanboy

[karnal]

Family Guy quote:

"No...Hot..."

Of course, I'm at work and took your comment literally (got "burned")..

nevermind. I need more sleep.

Re:Back to Intel Fanboy

[Loki_1929]

"Now... given this kind of statistics, as sad as it may sound I'd say I am willing to pay anything for an Intel just to avoid the headaches."

You apparently are unaware of the headaches others have suffered at the hands of the almighty Intel. Here's a refresher:

I Win.

Re:Back to Intel Fanboy

[mrm677]

I agree. AMD makes good processors but I don't think they have been as forgiving to newbie system builders. My experience is with older Athlon chips...hopefully they've improved since but I'll stick with Intel for now.

• 
  Cooling is/was more important, especially for the older T-birds.
  
• 
  You have to be more careful not to crack the chip when putting the heatsink on.
  
• 
  AMD has had software issues. For example, Win2K had to be patched to SP1 because AMD messed up AGP coherence. AMD also never told the Linux developers about this same problem causing numerous people to have system crashes when using nVidia cards/drivers under linux. You'd think they would have sent a fricken e-mail to the kernel-dev lit.
  
• 
  Socket A motherboards always seemed so damn sensitive to power supplies. I've got a Dell 4100, with a PIII 866MHz, filled with 4 drives and a Geforce4 Ti4400. All of this runs fine on the 200Watt power supply. Now maybe my problems with Athlons were related to shoddy motherboards, but I frequently ran into PS issues. Seems that good, high-quality Antec caliber supplies are needed. Yet I've built P4 boxes with 300W generics and have no problems.
  

Just my 2 cents. Competition is good and I know several people who have had _no_ problems with AMD.

Re:Back to Intel Fanboy

[Mindcry]

i've had about 20 procs (amd/intel, one cyrix 333, one via 1.3ghz), and none of them have died on me... i even got a working 8086 kaypro 16 sitting around... in fact the only computer parts that have given me trouble are power supplies (when i put too much crap on two low a wattage) and ram (i've had a fair number of bad sticks)...

all cpus use stock cooling solutions and are not OC'd.

Re:Back to Intel Fanboy

[Hoser McMoose]

Cooling is/was more important, especially for the older T-birds.

Cooling is VERY important for all current processors. It all becomes relative. When the Thunderbird was current, it used quite a bit more power than the PIII that it competed against, so cooling was very important for that chip. Now, the ~70W that the T-Bird used is not at all abnormal, this is the same basic power consumption of an AthlonXP (Barton or Thoroughbred), Athlon64/Opteron, P4 "Northwood" or even an IBM PowerPC 970 (aka Apple G5). The Intel P4 "Prescott" chips are a bit hotter, so for the moment people are talking a lot about how hot those chips get, but give it another year or two and 100W TDP probably will not be at all abnormal for processors.

You have to be more careful not to crack the chip when putting the heatsink on.

I've always wondered how in the hell anyone managed to crack their chips putting heatsinks on. I've put heatsinks on a fair number of AthlonXP chips and never even see a possible way to crack the core. Do people usually install heatsinks using a hammer or something?! While I do like the newer retention mechanism used on Intel P4 chips or AMD Athlon64/Opteron chips, it's really NOT difficult at all to put a heatsink on an AthlonXP processor.

AMD has had software issues. For example, Win2K had to be patched to SP1 because AMD messed up AGP coherence. AMD also never told the Linux developers about this same problem causing numerous people to have system crashes when using nVidia cards/drivers under linux. You'd think they would have sent a fricken e-mail to the kernel-dev lit.

You need to do a bit more reading on that problem, it was actually caused by nVidia's drivers doing some really stupid things (ATI does/did the same stupid things, as did most other video card vendors). It was only a matter of sheer luck that the problem DIDN'T affect Intel chips in the sameway.

Re:Back to Intel Fanboy

[Joseph Lam]

Actually, one can start to get a statistically significant result at around as little as 30 data points depending on a few factors.

Yes, but doesn't it require a random set of samples? This assumption no longer holds when people typically buy CPUs from their favorite retailers, who in turn get stuff from their favorite...

And regarding the original post about having 3 AMD CPUs dead, I'd highly suspect it's due not to the chip themselves (e.g. motherboard, cooling, bad retailer...). But anyway we can never draw any meaningful conclusion with that kind 'statistics'.

Re:Back to Intel Fanboy

[mog007]

Isn't it obvious? Grandparent poster only had 4 hsf available, and none of them fit the AMD sockets.

Re:Back to Intel Fanboy

[Anonymous Coward]

My AMD processor computers have given me trouble. My AMD machines locked up more frequently than my intel machines.

Re:Back to Intel Fanboy

[comedian23]

It could definately be a bad batch or it could also be a bad mobo mixed with a Homer Simpson moment.
*Installs CPU*

*Turns on PC*

ZAP!! *Sparks fly*

"Doouu CPU is bad"

*Install CPU*

*Turns on PC*

ZAP!! *Sparks fly*

"Doouu CPU is bad"

etc, etc, etc.

-Comedian

Re:Back to Intel Fanboy

[Anonymous Coward]

Hundreds of thousands?? That's complete crackhouse stuff. By numbers alone the odds that intel is more stable than amd is about 128:1 there. What really matters is that the methodology is right (are these really random processors? are you sure there's nothing else in common with the failed systems? exactly 7 processors were tested etc.)

Re:Back to Intel Fanboy

[NerveGas]

Out of the 100 or so CPU's that I've bought over the last ten years, I've had exactly *one* go bad on me. (FWIW, that was a Cyrix.)

Now, I've received some bad ones - but we're talking about a working CPU turning into a non-working CPU. That's actually a pretty rare thing - not just in my experience, but also in the experiences of my associates who have dealt in much larger numbers of CPU's than I have.

If you're had that many of your chips go bad, either (a) YOU are doing something wrong, (b) you're buying from someone involved with something shady, (c) there's something extremely abnormal in your environment, or (d) you're an amazing statistical anomoly, and should go out and buy a lottery ticket immediately. :)

steve

New hammers on the way.

[eddy]

Some info here. SSE3 is the big thing.

Re:New hammers on the way.

[Dominic_Mazzoni]

Some info here. SSE3 is the big thing.

I doubt SSE3 will make a difference for very many applications. A quick overview of x86 vector instructions:

MMX: primarily vector integer instructions on 64-bit registers; main flaw is that they use the floating-point registers, so you can't mix MMX and FP code. Biggest win for image processing, which is usually 8-bit data and perfect for MMX.
3Dnow: adds vector floating-point instructions on 64-bit registers (introduced by AMD).
SSE: primarily vector floating-point instructions on 128-bit registers
SSE2: vector double-precision floating-point instructions on 128-bit registers
SSE3: only 13 new instructions (compared to dozens for each of the other extensions). no new vector types, just some complex arithmetic, horizontal arithmetic operations, and unaligned data loading.

L3 cache

[Anonymous Coward]

I thought the very definition of L3 cache was off die. If it is on die, wouldn't it be L2 cache, unless is does not run at core CPU speed?

Re:L3 cache

[castlec]

the very definition of l3 cache, is level 3 cache, hence the name, l3. do you not remember when l2 cache used to be on the mobo?

Re:L3 cache

[Have Blue]

L3 is just the next larger cache after L2. Where it is in the physical circuitry is an implementation detail.

Re:L3 cache

[Anonymous Coward]

so does it run at full CPU speed?

Re:L3 cache

[Hoser McMoose]

Yes. It has a narrower bus to the processor when compared to L2 cache and has higher latency, but it's still on-die and running at full core speed.

Re:L3 cache

[fitten]

L3 cache is a generic term. You can think of the "L" as meaning "Level". L3 is 3 "steps" away from the CPU registers/core. Think of it this way (although it isn't necessarily implemented this way as there are optimizations that can be done to make this faster):

The CPU needs a value. If it isn't in a register, look in L1 cache first. If not in L1 cache, look in L2 cache. If not in L2 cache, look in L3 cache (etc through as many L* cache as you want). If not in L* cache, go ahead and fetch it from main memory.

Whether or not L3 cache runs at the core clock frequency or whether it is on die or not is completely dependent upon the implementation (which CPU are you talking about?).

For the P4 Xeons, the L3 is on the die and runs at core frequency (just has a few clocks of latency to access it).

Re:Suck...

[scharkalvin]

Why can't AMD make a Dual AMD64?
They DO. It's called an Opterion!
Guess you are refering to the Athlon64 which is a one way processor (just like the Pentium 4).

BTW the Opterion is made in 3 flavors, the 100 series is a SINGLE way cpu with NO smp support. It's very nearly the same as the Athlon64-FX. The Opterion 200 series is a 2 way (2 cpus), and the Opterion 800 series supports up to 8 cpus. AMD dropped plans for a 400 series, but you can use the 800 series chips to build servers with 1-8 cpu's.

Basically... if you have more then 2 use AMD

[Fallen Kell]

The jist of the whole thing is that Intel's achitecture has a huge bottleneck in its FSB. All the processors share the same FSB and quickly max it out if there are more then 2 processors. So anyone building or buying systems with more then 2 processors will get much better performance out of an AMD opteron system then an Intel.

Re:Basically... if you have more then 2 use AMD

[Anonymous Coward]

Exactly, Opteron arch. scales very nicely.

Re:Basically... if you have more then 2 use AMD

[hackstraw]

The jist of the whole thing is that Intel's achitecture has a huge bottleneck in its FSB. All the processors share the same FSB and quickly max it out if there are more then 2 processors. So anyone building or buying systems with more then 2 processors will get much better performance out of an AMD opteron system then an Intel.

Do you really consider 10->13% as "much better performance"? I must be a god, because I rewrote some queries on a database a few years ago and I got up to 400% performance without spending a dime on hardware.

Also, most people that go beyond 2 procs go with a real computer, and not a PC.

Re:Basically... if you have more then 2 use AMD

[MikeBabcock]

In other words, the switched bus technology they licensed from DEC payed off. This was what they hyped years ago as one of their main advantages over Intel in the not-so-distant (at the time) future.

Re:Basically... if you have more then 2 use AMD

[ELiTeUI]

No, actually, the Opteron's use a newly designed bus, namely HyperTransport (www.hypertransport.org). It was designed primarily by AMD. It is not a switched bus, it is a fabric architecture.

just so you know.

ELiTeUI

Re:Basically... if you have more then 2 use AMD

[ELiTeUI]

Most people going beyond 2 procs go with a 'real computer' because of the limitations of the intel bus/memory architecture(s).

The point is that opteron has shed most of these shitty limitations.

ELiTeUI

Re:Basically... if you have more then 2 use AMD

[Anonymous Coward]

I think IBM, HP and NEC all offer systems using Intel CPU's with their own cell based motherboards to support up to 64 CPU's. HyperTransport is going to make life good for AMD and bad for Intel. It makes me wonder why we're even bothering with Bus anymore. Considering AMD has hypertransport, apps that are I/O heavy will continue to see AMD doing better than Intel.

Re:Basically... if you have more then 2 use AMD

[brokenbeaker]

even with two processors, AMD and Intel are neck and neck in performance, but the former is cheaper.

Re:Basically... if you have more then 2 use AMD

[pla]

Also, most people that go beyond 2 procs go with a real computer, and not a PC.

True, for a reason totally unrelated to architectural constraints...

I can get a 2-way motherboard for under $100. A 4-way motherboard starts at over $1500. I can't even find a price for an 8-way PC-style system, but I take that as a sign that they don't come cheap.

Now, I currently run a dual P3/933 as my main machine (yeah, about time for an upgrade, but only because I can get a much better replacement dirt cheap, it still runs just fine). The motherboard cost me $120, the CPUs $55 each at the time - Less than half the cost of the board itself. If I could have found a decently priced 4-way motherboard, I'd have used that in a heartbeat for a mere $110 more in CPUs (yes, I realize that the non-Xeon P3 didn't support a quad config, but that actually works to support my point, as I'll explain).

Commodity-level 4-way, I see, as the big advantage that the Opteron may finally give us (apart from greater than 4GB of memory per process). To get the top-of-the-line chip usually costs 3-4x last year's best offerings, for only 10-25% better performance. For the same price (in terms of just the CPU), you can get a dual or even quad system. Now that a PC-level chip exists that does support 4-way and above, I believe we may finally see reasonably priced motherboards supporting that many CPUs.

The "big boys" will always spring for their top-of-the-line 8-way systems. For the rest of us, a quad system based on last year's CPUs would make for a truly sweet machine, and (if the 4-way motherboards become reasonably available), wouldn't break the bank.

Re:Basically... if you have more then 2 use AMD

[imsabbel]

No, the ev6 bus was a miserable failure as far as multiprocessing goes. Beeing point to point, there were a lot of problems with cache coherence (not really problems in the sense of "it doesnt work", but in the sense of "damn. why cant we get faster").
Plus the need for a few hundred pins per cpu on the northbridge made it impossible do to more than 2 cpus anyway (and even with only dual, the 760MP northbridge was a >1000 pin monster).

Re:Basically... if you have more then 2 use AMD

[dcam]

If cost isn't a factor. Anand makes the point that the Opterons are less that half the price of the Xeons.

Re:Basically... if you have more then 2 use AMD

[comedian23]

>Do you really consider 10->13% as "much better performance"?

In anything where you compare speed 13% is a big difference. Imagine if Deion Sanders had only been able to run a 4.4 40 instead of a 3.8, he would be at best an "decent" CB? Imagine if you can cut processing on huge data sets down from 5 days to a little over 4( so you can actually do something with the data on Friday before you leave). Or so you can buy 1 less machine to do the same work and save a good chunk of money.

It it was 1-2% I might agree with you, but 10-13% is pretty significant.

Not that any of that diminishes you though. Sounds like you did some top notch work on that DB. I certainly wish we had someone with your DB knowledge here.

-Comedian

Re:Basically... if you have more then 2 use AMD

[Fallen Kell]

Notice that the 10-13% increase is in the 2 way system. Now that we have maxed out the FSB, above the 2 way system (i.e. 4 way or greater) since communications have already maxed out, you will not even see an increase in performance with the addition of more CPU's, as there simply isn't a way to get them the data for them to actually process anything. But on the other hand, the AMD opteron's will continue to gain performance with the addition of more CPU's and will gain more and more performance. The benchmarks were not completed in the 4 way systems simply because Intel was approx ~70-80% slower then the AMD. And at 8 way it would approximate the AMD to be 300-400% faster then the Intel simply because at that point Intel is not gaining anything for the additional processors beyond 2 when it maxed out the FSB while AMD is still getting 50-65% of the processing power of each new CPU that it adds on.

Re:Why benchmark games?

[Anonymous Coward]

It is bad enough that he didn't read the article but even after everybody else pointed out that the original poster was wrong you still posted your equally uninformed response. He is a troll but I am not sure what that makes you.

Price vs. Performance

[maharito]

I attend a university that is currently building a beowulf cluster, and when it came down to making a decision, the deciding factor was price/performance ratio. While it may make sense for enterprises to go with the Xeon, the Opteron is a clear winner, in my mind, when money is an object. Of course, if you have the money to burn, the Xeon may seem to be the more obvious choice.

Re:Price vs. Performance

[Pingular]

I attend a university that is currently building a beowulf cluster, and when it came down to making a decision, the deciding factor was price/performance ratio. While it may make sense for enterprises to go with the Xeon, the Opteron is a clear winner, in my mind, when money is an object. Of course, if you have the money to burn, the Xeon may seem to be the more obvious choice.
Even if someone has money to burn, wouldn't it be better to get more performance anyway?

Re:Price vs. Performance

[Mr Guy]

Boy I'd like to see a beowu... oh.

Re:Price vs. Performance

[Mateito]

Imagine a... Oh. Wait. C-mon... somebody was going to say it.

Re:Price vs. Performance

[jazman_777]

I attend a university that is currently building a beowulf cluster...

Whoa, slow down, pardner! You're only supposed to _imagine_ these things, not really build them.

Re:Price vs. Performance

[Rich0]

Of course, if you have the money to burn, the Xeon may seem to be the more obvious choice.

No, if you have money to burn, then more Opterons for the same price as the Xeon is the more obvious choice. I mean really, if $x can buy you 1000 Opterons or 500 Xeons, and they are pretty comparable in performance, then wouldn't it make sense to get the 1000 Opterons, or just get 500 and save half the cost?

Re:Price vs. Performance

[haggar]

Of course, if you have the money to burn, the Xeon may seem to be the more obvious choice.

"Of course"? Why?

Re:Price vs. Performance

[Eneff]

Because it's not a s simple as X vs 2X computers. The chip, while significant, is not the total cost of the computer. Xeon 2x motherboards, for example, run 200 dollars less for equivalent Tyan MBs. (from looking at Newegg, anyway)

Because even if the chip is a significant portion of the cost of building the computer, it is only a small fraction of the total cost over the useful lifetime of the cluster.

Because one has to benchmark for one's own problem set. It's possible that one set of instructions are better optimized for Xeons.

Because the fewer number of nodes in a cluster, the more efficent each individual node is. A small performance increase may be substantial enough to require fewer nodes, bringing numbers into line.

Because if it's big enough, Intel might throw in a few days with an engineer to sweeten the deal. (But then again, so may AMD.)

Numbers arguments get too complex to make such an important decision a no-brainer.

Re:Price vs. Performance

[cilix]

You seem to be forgetting that there are other factors to consider when working out price. A significant part of the price of running a large cluster is cooling. You get a couple of thousand procs in a room, and airconditioning becomes challenging. I don't know how hot opterons are in relation to xeons, but it's a very important factor that needs to be considered.

Re:Price vs. Performance

[Anonymous Coward]

No, if you have money to burn, the obvious thing is to videotape it.

Back to trolling.

[Anonymous Coward]

Maybe some adult should come and take your toys way from you if you can't play nicely.

I've never heard of a CPU just breaking down, there's always some kind of abuse (running without heatsink -- which is no longer an issue on the processors on topic -- or they've been 'chipped' due to cluesless users (redundant) who don't know how to install a heatsink.

memory controllerS?

[Bender Unit 22]

But these days days with all the virtualization getting hot(vmware etc), a server architecture with a single memory bus/controller is getting old.
I'd like to see some test on servers like the IBM x445 with NUMA.

Re:memory controllerS?

[Anonymous Coward]

Did you even bother to read the article? Opteron scales better because it has one memory controller per CPU.

Idjit.

Re:memory controllerS?

[ysachlandil]

Maybe you need to read the article before posting, or the Opteron technical docs for that matter. The x445 is a Xeon based server with some nasty hacks to get the Xeons to play nice with the multiple memory controllers.

The Opteron has a built in memory controller, and when you build multiprocessor systems, each CPU gets its own memory. The HyperTransport chip interconnects make sure all the CPU's can reach each other's memory, and has been designed from the ground up to be as low latency as possible.

Read this presentation before you try to convince us Xeons in a x445 are a better NUMA solution:

http://computing-colloquia.web.cern.ch/computing -c olloquia/talks/Opteron 09.11.03.ppt

check out pages 22 and 23 for a nice comparison between legacy (read Intel) and Opteron multiprocessor memory architecture...

--Blerik

Re:Suck...

[Fulcrum of Evil]

Why can't AMD make a Dual AMD64?

They do - it's called an Opteron 2xx.

Dual AMD64 is made, and works great.

[Anonymous Coward]

They call it the "Opteron".
Let's use the Athlon64-FX as a base.
Relative to that, the other chips are:

0. self
1. Plain Athlon64 (not FX) has narrow memory.
2. Opteron for 2-way systems
3. Opteron for 4-way systems
4. Opteron for 8-way systems

So that's 5 chip models.

The Usual Problem

[Sloppy]

We've seen this same type of benchmark over and over. It wasn't interesting then, and it's not interesting now.

The tests in this article, involved running the same exact binaries (out-of-the-box Microsoft 386 stuff) on both types of CPUs, rather than the code being compiled to run natively. The Opterons were fighting with one hand tied behind their backs.

In other words, this benchmark is mainly only of interest to Microsofties. If that's what you run, then fine, the article may be useful to you and you may get something out of reading it.

If you are trying to maximize speed, though, then the software contraints that this test took place under, are totally contrary to what you'd actually be doing (running code that is appropriate for the hardware).

BTW, another weird thing I noticed about this article: these guys use flash for static images of bar graphs. WTF? Anandtech, your w3b d3$1gn3rz R S0 31337!!!1

Re:The Usual Problem

[Ancient_Hacker]

What a weird way to benchmark a 64-bit CPU-- with 32-bit software, 32-bit OS, and a disk-intensive test. Perhaps a more releveant test would be to use 64-bit apps ona 64-bit OS, and tests that tend to exercise the element in question, the CPU. Methinks the test designers were either really, really dumb, or they were aiming to minimize the perceived speed differences between the contenders.... (or they're both).

Re:The Usual Problem

[Anonymous Coward]

Bah. In all likelyhood, even Unix/Linux users aren't going to be custom compiling their Inventory System for optimal performance.

You're right that these benchmarks only cover RDBMS-based client-server shrinkwrap apps, but that's a pretty common case. When Oracle and MS ship their DB Servers compiled for x86-64, the tests should be run again, but until then business customers will need to work with what they've got.

Claiming that this is "mainly only of interest to Microsofties" is borderline flamebait, especially since one of the marketing points of the Opteron is excellent performance with your existing software.

Re:The Usual Problem

[hackstraw]

The Opterons were fighting with one hand tied behind their backs.

I thought that one of the selling points of the opteron was that it could run 32bit and 64bit at the same time and at full clock speed (ie, no emulation, etc).

Re:The Usual Problem

[Sloppy]

I thought that one of the selling points of the opteron was that it could run 32bit and 64bit at the same time and at full clock speed (ie, no emulation, etc).

The selling point, aimed at the mainstreamers locked into the 386 legacy, is that it runs 386 code quite well. So when you replace your old computer with a new one that has an AMD chip in it, it'll still be an upgrade rather than a downgrade.

But that doesn't mean it can't be faster. To the large mainstream market, the fact that it isn't as fast as it could be, isn't a big deal. If it were, they would have all bought Itaniums instead of Pentium 3s. Heck, they would have all bought Alphas in the early 1990s. But obviously they didn't do that, because performance was a tertiary value, compared to price and compatibility.

But when you get to the segment of the market who values performance more highly than binary compatibility with the old legacy, things change. Throw in the memory contention issues that start to pop up with multi-CPU sytems, and running 386 code becomes even less important. I mean, one of the first things that'll pop into your head will be, "Uh, if memory is so slow, why don't we use some of the extra general-purpose registers that x86-64 has to offer? Screw 386 compatibility, let's compile our code to run as fast as it can." Seems to me, these are exactly the kind of people who would be trying to build fast systems, getting boners over one processor being 3% faster than its competitor in some benchmark, etc.

Now the question is: which of those markets, actually cares about performance benchmarks of Opteron vs Xeon, and which of those markets, is looking at pricewatch comparisons of Duron or Athlon64 vs Celeron?

This Anandtech article is targeting .. who? Neither. People who buy high-end hardware and then run low-end legacy software on it, that's who. In other words, idiots.

Re:The Usual Problem

[NerveGas]

Actually, I really imagine that the benefits of an Opteron can only get better on better operating systems.

M$' garbage is made for the traditional shared-bus style of design, which the Xeon uses, and isn't designed to take full advantage of the NUMA architecture of the Opterons. With a better OS, the Opteron is likely to take even more of a lead.

M$' garbage *does* have support for hyperthreading, giving the Xeons a boost. However, it does *not* appear that the version they used had support to run in 64-bit mode on the Opteron. Even though the bitness isn't going to speed it up, 64-bit mode has more registers, which would benefit the opteron in these results. Again, with a better OS, the Opteron would be likely to take a bigger lead.

So, the benchmarks aren't perfect by any means, but they're not completely useless, either. If nothing else, it's a good way to rub Intel's nose in the fact that their overpriced chips are being beaten by bargain-priced underdogs, even when the cards are stacked in Intel's favor.

steve

Re:The Usual Problem

[NerveGas]

Anandtech seems to do all of their business with Microsoft products, they don't really seem to be completely aware (or competent) with Unix or any Unix-like operating system. Instead of malice, I'd suggest ignorance as a more probably cause. : )

steve

Re:The Usual Problem

[Hoser McMoose]

Perhaps a better test would be for them to figure out what people actually did with their servers and tried to duplicate it as closely as possible.

Ohh wait, that's exactly what they did.

Re:The Usual Problem

[Anonymous Coward]

Actually, Win2003 does have NUMA support, but I have no idea if it's used (in Windows or Linux) for small Opteron systems.

Re:The Usual Problem

[WNight]

You don't need to recompile your inventory system, it's probably some light-weight front-end for the stored procedures in the DB. The DB engine could be recompiled though.

It would be interesting to see this on a 64-bit Linux with PostgreSQL recompiled.

For every staid fortune 500 company there are tens or hundreds of smaller companies like ISPs and web stores who run everything in Linux and are small and nimble enough to recompile parts of their software to take advantage of this.

Re:The Usual Problem

[WNight]

Sure, it runs 32b code at full speed, but you're ignoring the benefits of being able to run 64b code.

For something that should be fairly system agnostic (an SQL database) you could consider going to a 64b Linux and 64b PostgreSQL.

Article on one page

[mulle]

Full text

So...

[Anonymous Coward]

So basically you want Opterons?

Re:Suck...

[Hektor_Troy]

It's called an Opterion!

Well, you can get dual Opterions, that is correct, but they seem to be contact lenses. Not sure you're like to plug that in to a motherboard ... at least not while wearing them.

Re:Suck...

[garompa]

Opterion ? you meant Opteron right ?

2 is better than 1

[ssbljk]

Results were a little varied as 4-way Opteron systems seemed to fare the best, although dual Xeon configurations almost always beat dual Opterons.

well, I prefere 2 machines with 1 Opteron each than 1 machine with 2 Xeons :)

Hmm

[Trailer Trash]

...4-way Opteron systems seemed to fair the best...

Did they fare well, also?

Re:Hmm

[sik0fewl]

Obviously not or he would've exlicitly stated it in his article.

Re:Hmm

[Anonymous Coward]

They fared fairly well. But to be fair, let's wait until the tech fair to see a head-to-head comparison. Too bad I can barely afford the bus fare... It's not fair.

Re:Hmm

[Anonymous Coward]

Exlicitly... Is that something that used to be licit, but is now illicit, hence exlicit?

Umm.. wrong.

[Egekrusher2K]

There were but a very few benchmarks that the Xeons beat the Opterons on in the 2 way configuration. And even those were by a very small margin. And in the 4 way configs? It was a slaughter.

Re:Umm.. wrong.

[Anonymous Coward]

In the real world when the 2-way Xeon beats the 2-way Opetron a smart admin would simply buy 2 more opterons and go 4-way instead (read: AMD 1/2 the price of Intel). That way, the low cost itself actually means better performance. So reardless of what the benchmarks said, Opetrons slaughtered the Xeons even in the 2-way benchmark (in a non-technical way). The submitter of the article saw what he wanted to see and apparently didn't catch that. It sounds weird, but that's price/performance scaling in a nutshell. Even if you went 4-way+ on Intel, the FSB bottleneck would kick in, making Opetrons even more beautiful.

It's amazing how well AMD has been doing lately.

Why is this news, seriously?

[Anonymous Coward]

There have been *SEVERAL* benchmark comparisons already, for MONTHS, and this one came to the same conclusion as the others. Nice to see that Anandtech is months behind the competition and their contribution is nil.

Re:Why is this news, seriously?

[tangent3]

Anandtech had an Opteron vs Xeon test earlier too, AMD Opteron 248 vs. Intel Xeon 2.8: 2-way Web Servers go Head to Head where the Opteron trashes the Xeon handily. I guess that was more focused towards web serving, and now that Anandtech intended to replace their forums database server, they naturally based their latest test "AMD Opteron vs. Intel Xeon: Database Performance Shootout"on database performance.

"In a 4-way configuration AMD's Opteron cannot be beat, and thus it is our choice for the basis for our new Forums database server. We'll be documenting that upgrade in a separate article so stay tuned."

Re:Why is this news, seriously?

[Egekrusher2K]

Because of the new release of the 4MB L3 cache Xeon, you jackass, and because of the added support of AMDs 64 bit extension support.

-5, Clueless

[Gothmolly]

Firstly, Anandtech uses flash for its images so that people w/o the plugin can't see the data. This forces you to install it, so that you can see their OTHER Flash pieces... ads.
Secondly, you are not going to get MS to recompile an MS-SQL for Opteron. You're not going to get IBM to support a Linux installation, after you've rolled your own ueber-NUMA-patch-level-42 kernel.
The test was clear - out of the box, plug in servers, load OS, load app, run benchmark.
And the outcome was clear, the Opteron architecture is vastly superior, both performance and price-wise.
The MHz myth is over, at least in Slashdot and Anandtech circles.

Re:-5, Clueless

[MikeBabcock]

1) Microsoft is a major Opteron supporter; they had a freely downloadable Opteron Windows XP beta available for some time now that I have an ISO of here.

2) IBM would probably support uber-numa patched kernels as you put it, since they are one of the main proponents of Linux-on-massively-parallel-supercomputers anyway.

Do some research.

Re:-5, Clueless

[FuzzyBad-Mofo]

Anandtech uses flash for its images so that people w/o the plugin can't see the data.

In my opinion that's a pretty lame tactic. I don't mind banner ads, but animated, noisy Flash advertisements are where I draw the line. Good thing there's Firefox's Flash click-to-view plugin!

Re:-5, Clueless

[Sloppy]

You're not going to get IBM to support a Linux installation, after you've rolled your own ueber-NUMA-patch-level-42 kernel.

Wouldn't IBM be happy to give you a kernel with the NUMA support, since they're the ones who stole it from SC-- err, I mean -- wrote the NUMA stuff?

;-)

Re:-5, Clueless

[cens0r]

Secondly, you are not going to get MS to recompile an MS-SQL for Opteron.

Certainly not for a benchmark test. But since intel and AMD are both supporting x86-64 it seems, I gurantee their will be a x86-64 MS-SQL in the near future.

Re:-5, Clueless

[brewt]

I myself prefer the newer, better FlashBlock Mozilla/Firefox plugin.

Re:-5, Clueless

[Anonymous Coward]

I an neutral in this argument, but:

http://www.theinquirer.net/?article=8815

Initial investigation points to one thing MS is doing that you claim they aren't.

Going for the old percents... you have 2 issues.. part of #2 seem wrong, so all of #2 is wrong. That leaves you with at most 50%... which is below my threshold for belief.

Re:-5, Clueless

[Junta]

Hmm... moderated yourself nicely.

IBM supports native x86_64 distros (SLES8/AMD64 now, and looks like RHEL3/AMD64). So IBM supports you running 64-bit today.

Windows is working towards 64-bit for Opteron, so yes, MS is recompiling their crap for Opteron.

Simply because there is a logical reason for images being flash makes it no less annoying as hell.

Opteron is a fascinating platform, and very cool, *especially* with respect to 64 bit computing.

Re:-5, Clueless

[teg]

Firstly, Anandtech uses flash for its images so that people w/o the plugin can't see the data. This forces you to install it, so that you can see their OTHER Flash pieces... ads.

Get flashclick. You can then install the flash plugin, and you'll have to click on the flash areas to download/run it. Now you can look at the sites which really need it, while still avoiding the flash ads. Very, very nice.

Re:-5, Clueless

[Anonymous Coward]

you are not going to get MS to recompile an MS-SQL for Opteron. You're not going to get IBM to support a Linux installation, after you've rolled your own ueber-NUMA-patch-level-42 kernel.

Good point! Wasn't this the whole point of the Opteron in the first place? The fact that it can run (and run quite well) the existing code-base is its very reason for being!

Re:Suck...

[Wrexen]

Here at /. we have a long history of mispelling AMD products, going all the way back to the Athalon!

Conclusion...

[ERJ]

Anand seems to conclude something a bit different then the submitter:

The comparison we've made here is a very important one; it identifies Intel's strengths and their weaknesses with Xeon, and it crowns Opteron a clear multiprocessor winner. An area that we didn't touch on is cost, which is where AMD truly shines. The Opteron 848 processors we tested are around 1/2 the price of Intel's 2MB L3 Xeon MPs and we have not seen retail data on how expensive the 4MB parts will be.

In a 4-way configuration AMD's Opteron cannot be beat, and thus it is our choice for the basis for our new Forums database server.

Re:Suck...

[Anonymous Coward]

Because AMD64 has only ONE Hyper Transport link that gets used to connect the southbridge.

Opteron 2xx have 2 HT links: 1 gets used for SB, 2x1 to connect to each other, and 1 ist left for either another SB or left dangling - I believe you could chain more than 2 of them, but accesses across all the chain would be slllooowwwww....

Opteron 8xx have 3 HT links, and with more than 2 processors you may want fewer SBs than processors for more connectivity between processors.

Question For Geeks Intel Hyperthreading

[Anonymous Coward]

Is Intels hyperthreading tech BS or not ? I am thinking about building another system and am considering Intel & AMD. I have not really found anything about hyperthreading that was not just marketing crap from Intel.

Links and replies appreciated

Re:Question For Geeks Intel Hyperthreading

[bluGill]

Hyperthreading is interesting. According to the article it is worth about a 5% boost overall. Note the overall. If you are running one single threaded application it won't help. If you run lots of things at once it will give a 5% boost.

I don't think you will notice it. However if you do decide on the P4 you would be stupid to not enable it, it costs nothing and just works. There are lots of factors in choosing your system, you can ignore hyperthreading because the other factors are so much stronger.

Re:Question For Geeks Intel Hyperthreading

[Hoser McMoose]

Definitely not BS, though whether or not it's useful depends heavily on your application.

The idea behind hyperthreading is that the P4's long pipeline will often stall with only a single thread going through. With hyperthreading you run two threads at once, so when one thread stalls you just start up the other thread and go with that one for a while. In a way it's almost like a poor-mans dual-processor system, giving you two logical processors on a single chip.

Now, obviously there are a few things to consider here. First off, if ALL of your processing is being done in a single thread then you aren't going to see any benefit to hyperthreading, and in fact the extra overhead might even make things a bit slower (usually only 1-2% slower).

Games almost always do all their major processing in a single thread. Even if they have extra threads hanging around, you almost always spend 99%+ of your time in a single thread. For this reason, games see virtually no benefit to hyperthreading (they don't see much/any benefit from dual-processor setups either).

On the other end of the spectrum, some applications see up to a 25% performance boost when hyperthreading is enabled. The tests I've seen show the biggest improvement have been things like Photoshop and rendering applications. Some server applications should benefit as well.

The other boost that hyperthreading gives you, like with a real dual-processor setup, is that it makes multitasking a bit "snappier". This is by no means a night-and-day difference here, but it is there.

Re:Question For Geeks Intel Hyperthreading

[Anonymous Coward]

AMD opted not to pursue Hyperthreading (psuedo dual processing), and will introduce a full DUAL CORE Opteron within the next year (after the die shrinks again- basically it will be 2 opterons on a single core, sharing the same L2 cache- part of the beauty of HyperTransport design).

I'll wager the performance increase will be greater than 5% overall (but less than 100 because of cache contention).

If Intel is still on the shared bus/northbridge, I think they'll be hurting.

Re:Question For Geeks Intel Hyperthreading

[TheLink]

The reasons why some games don't get any benefit from hyperthreading is because the programmers have optimized the code so there won't be spare ALU/integer and floating point units to do other stuff.

Programs get benefit if the code isn't as optimized and there are spare units unused by one process thread which can be used by another thread.

It seems to me that unoptimized code is more likely to benefit from hyperthreading than optimized code. Trouble is, code is often optimized when performance is a high priority.

Very often hyperthreading makes things significantly slower. Seems too much like a gimmick.

I wonder if antivirus software support hyperthreading - e.g. work better with it enabled. That would be interesting - a significant portion of personal computers in the world have antivirus software installed doing real time scans and thus using up significant CPU - even if the AV software is not up to date in most machines (and thus useless) it still imposes a CPU load.

I haven't seen CPU oriented benchmarks that involve antivirus software - e.g. is P4 w or w/o HT better than Athlon or Athlon64 for AV? Antivirus software is pretty much a "real world" requirement in many places.

Re:Why benchmark games?

[nolife]

I'll soon be finishing up my extensive long term testing of an Apollo 735 HP-UX Unix Workstation with a 125Mhz PA-7200 PA-RISC processor. I'll post the results for you if you are in the market for one of these. You can still find them on Ebay for about $5. ;)

What kind of servers are they using?

[Anonymous Coward]

The lack of detailed hardware specifications disturbs me. What servers are they using? Vanilla clone quad proc servers? If so, odds are both the AMD and Intel versions could do better using hardware actually optimized to take advantage of them. I'd like to see HP servers running both chips put to the same tests and see how they perform.

Re:What kind of servers are they using?

[NerveGas]

I highly doubt that HP is making their own chipsets, or coming up with any other improvements that would give them a significant *hardware* advantage over the others. And with the Opteron being so self-contained, it would be even harder. With the memory controller being integrated, as long as you have the HyperTransport lines wired up, then most of the performance difference will come from memory timings and the like.

Out of the quad-Opteron arena, the Tyan is (as you imply) a complete joke, its memory design is completely, retardedly crippled. However, the vast majority of quad-Opteron boards are either AMD's design or NewIsys' design, and those two companies are betting the farm, so to speak, on the Opteron. I have a hard time believing that either of them are just lazily selling unoptimized boards.

Also, if anything, "vanilla" boards tend to be *overly* optimized, at the possible expense of stability. As an example, if you take a look at the *actual* FSB on a good number of boards, you'll find it measurably higher than what the BIOS screen told you it was.

steve

Apples, Oranges, and Efficiency

[Phaid]

It's interesting to notice that in these tests, the Opterons were clocked quite a bit slower and had a lot less cache than the Intel CPUs, yet performed comparably in 2-way and better in 4-way than the Intel chips.

The Opteron clocked at 2.2ghz with 1MB of cache was very close in 2-way performance with the Xeon 3.0 and 3.2 ghz each with 4 and 2 mb of cache respectively. The 1.8ghz Xeon compared well with the Xeon 2.8ghz with 2MB of cache. The Opterons were typically within 3% or so of their Intel counterparts in 2-way benchmarks and closer to 10% ahead in 4-way.

If nothing else, this says a lot about the efficiency of the Opteron's design. Less silicon, and more importantly for AMD, less expensive silicon, manages to achieve very close results.

Re:Apples, Oranges, and Efficiency

[Anonymous Coward]

athlons have always had better IPC than P4s.. its just that P4s can reach higher clock rates.

Anandtech isn't biased.

[brucmack]

The purpose of the test is not to test the memory, but to test the processors. Thus, they used the same memory in testing each processor configuration.

One of the purposes of the test was to show how the memory bandwidth bottleneck of the Xeons limits their effectiveness in 4-way configurations, which the Opterons do not have that problem. Doing this comparison with different memories would make things more complicated.

Additionally, you'll notice that Anand's final words recommend the Opteron for being at least equivalent and much cheaper than Xeon. This was also the selection process for their new forum servers, so you can bet that they aren't getting any kickback from Intel, or those would be Xeons.

If you still have doubts about the validity of Anandtech's testing, check out the benchmarks from their AMD vs. Intel web server test in December: http://www.anandtech.com/IT/showdoc.html?i=1935&p= 9. All on dual processor configurations. There is definitely no Intel bias in that test.

Really, I think some people ought to think before they flame like this. The benchmarks are showing the Opterons to be equivalent or faster in 2-way configurations and definitely faster in 4-way configurations, so what is there to complain about? The fact that Anandtech has consistently recommended AMD's processors just makes it doubly silly.

Re:Anandtech isn't biased.

[hackstraw]

The purpose of the test is not to test the memory, but to test the processors. Thus, they used the same memory in testing each processor configuration.

Then they should not have done an memory intensive and disk intensive benchmark like a database, now should they?

Re:Anandtech isn't biased.

[brokenbeaker]

I cannot agree with what you are saying - they are not testing chips, but architectures. in the 4 CPU configurations (and more), AMD scales much better because of the hyperlink and memory controllers on each die. if there is a memory configuration that AMD supports, and leads to better performance, why not use it?

Re:Anandtech isn't biased.

[Anonymous Coward]

" The purpose of the test is not to test the memory, but to test the processors. Thus, they used the same memory in testing each processor configuration."

The simple fact that the Opteron performs as well as it dose when crippled by slow memory should tell the Opteron naysayers something.

Signed Blueintheface

Re:Anandtech isn't biased.

[brucmack]

in the 4 CPU configurations (and more), AMD scales much better because of the hyperlink and memory controllers on each die. ...which becomes evident no matter what speed of memory one uses, and was indeed concluded in the article.

if there is a memory configuration that AMD supports, and leads to better performance, why not use it?

Because it makes it harder to directly compare the results of the AMD chips to the results of the Intel chips. The purpose of a benchmark like this is to say "all other things being equal, x is faster than y in situations {z}". This lets an interested IT professional look at the results and say, "ok, Opteron is faster than Xeon, now lets figure out what speed memory we can afford."

Honestly I don't see what the problem is here... it is obvious from the article that Opteron wins this contest overall, and anybody who is in the market to buy a dual or quad cpu server is probably going to know something about memory as well.

Re:Anandtech isn't biased.

[brucmack]

Then they should not have done an memory intensive and disk intensive benchmark like a database, now should they?

I think the point that they were trying to make was that Opteron performs much better than Xeon on memory regardless of the speeds. This is the whole reason why Opteron wins the benchmark, and this is made VERY clear in the article. They have published this information on Xeon's memory bottleneck many times.

I realize that what I wrote can be seen as meaning that the memory is irrevelant... I realize that it is not. However, using different memory in such a test brings in other variables, like memory timings.

By using the same memory on both architectures, they prove without a doubt that Opteron has a better memory architecture than Xeon. Had they used faster memory on the Opteron, the Opteron would have won, but it wouldn't be as clear whether it was because of the faster memory or the faster memory architecture.

Re:Anandtech isn't biased.

[brokenbeaker]

I just don't think we agree to what the testing is supposed to achieve. You think there should be only one difference that's being tested (unlikely that there is only one, given the architectural differences), while i think they should test "which system is fastest".

Like you say, there is no problem, just different approaches.

how are intel implementing 64 bit mode on the new.

[Anonymous Coward]

How are intel going to implement 64 bit modes on the new ia32e chips?

Is it going to be as badly as their implementation of 32 bit modes on the itanium?

Re:Pffft. These Intel vs. AMD flamewars are pointl

[ThisIsFred]

Going to go slightly off-topic here:

I'm an AMD lover, but it's my opinion that AMD is making a *huge* mistake with their desktop market. They only produced two marginal FSB400 processors with the "32-bit" Barton core, and then focused all their attention on the Athlon 64s. People who've made a choice in the past year to go with an AMD-compatible FSB400 mainboard are getting the shaft, and AMD is unwittingly forcing them to move to Intel during their next upgrade. Currently Intel's latest 3.0+ GHz offerings are spanking Athlon 64s in benchmarks with 32 bit applications. When users decide to do the next upgrade, they're going to say "hey, I have to replace my mainboard anyway", and they're going to go to Intel because it has more upgrade possibilities, is cheaper than the Athlon 64 for the same level of computing power, and currently performs better.

So this is more of a plea for AMD to extend the Athlon "32" line a bit further. Please AMD, don't prematurely kill off 32-bit Athlon chip development!

Re:Suck...

[mxf8bv]

Yeah, but the Opterions are exclusively available for dual unit use. Opterion marketing dept reports very little demand for Opterion-4 and Opterion-8 systems ;)

Not beat on the price front...

[Anonymous Coward]

So, purchase two dual opteron systems for the cost of that one dual xeon. Intel still has not gotten competitive on the prices. Until then only a fool would part with their cash on a xeon system.

Re:Suck...

[Deliveranc3]

AMD dropped plans for a 400 series, but you can use the 800 series chips to build servers with 1-8 cpu's.

Crackfiend.

Re:Suck...

[EinarH]

According to Anandtech it's possible to use the 1xx series in a dual setup with SMP support. This is not "validated for use" and nothing for business setups but still possible.

From this test:

With the Opteron, AMD introduced their first-ever 3-digit naming system for server/workstation CPUs. The first digit indicates whether the CPU was designed for 1-way, 2-way or 4- to 8-way operation. For example, the Opteron 100 series is only validated for use in uniprocessor configurations while the Opteron 200 series is validated for use in uni- and dual processor configurations. Finally, the Opteron 800 series is validated for use in up to 8 processor configurations.

Note that we used the phrased "validated for use" because there is very little stopping an Opteron 100 CPU from being used in a dual processor environment (an Opteron 100 is identical to an Opteron 200 and an Opteron 800). As far as we know at this point, AMD has not prevented CPUs from being used in configurations in which they weren't intended to be used. Although, it wouldn't be too hard for them to prevent it in the future if it becomes a problem.

Comparing Prices

[gbulmash]

4 AMD Opteron 248's at Newegg: $5876 ($1469 ea)
4 Xeons (@Intel's announced pricing): $14768 ($3692 ea)

Did the quad Xeon system outperform the quad Opteron by a factor of 2.5:1? No. In fact, in some cases, the quad Opteron outperformed the quad Xeon. The Xeon had advantages of hyperthreading, 4x as much cache, and a clock speed 800mhz higher than the Opteron, ans still got beat.

Clock speed may sell in the consumer market ("Me want bigger!"), but in the server market, Opterons getting better performance for half the price are going to win more and more converts.

- Greg

Re:Comparing Prices

[Anonymous Coward]

That's nice, but I'm not sure that difference will show up in the system price -- big vendors like Dell and IBM get huge discounts off Intel list.

Re:Comparing Prices

[tjw]

4 AMD Opteron 248's at Newegg: $5876 ($1469 ea) 4 Xeons (@Intel's announced pricing): $14768 ($3692 ea)

Actually, you'de need Opteron model 848 to do 4 way SMP. Obviously this is what you meant though, since you have the right price.

Re:Comparing Prices

[edgedmurasame]

At the prices they're going for, I'm surprised nobody's tried to find something to force the non-848's to do 4 way SMP.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[ckaminski]

Whilst I agree with you, and disagree with you, some of us (me) are hoping that AMD finds enough success in AMD64 to cripple a shitload of them as Athlons, and get out of the segregated 32bit cpu business.

But yes, I agree with you, AMD cannot neglect the desktop market, unless it makes AMD64 cheap enough that it can put them in all computers (which I think is their inevitable goal). Hell, once eMachines starts stocking them in Computer City, I think they'll have achieved it.

No Surpises Here

[ThisIsFred]

Just what I thought: AMD excels when the benchmarks include only moving around chunks of memory. Anything that involves more calculation and Intel has the edge, because their processor cores are running at almost 50 per cent faster. Nice to see AMD ahead of the game with the northbridge-on-die design. Don't worry, Intel will match that soon, since they've already surpassed/matched AMD with regard to FSB speed, on-chip cache and soon 64 bit extensions (all while keeping a healthy lead in clock speed).

Re:No Surpises Here

[BenjyD]

because their processor cores are running at almost 50 per cent faster

Because everyone knows more of them megahurtzy things is good,right?

Re:No Surpises Here

[Anonymous Coward]

unless Intel implements their own support for HyperTransport, uses AMD's 8000 chipset or comes up with their own switch fabric architecture, they're not going to beat AMD for 4,8,16 and 24 way systems. Pumping up Front side bus isn't going to fix the problem. Why do you think IBM, NEC and HP all use their own cell chipsets for their large 64 way servers, like HP SuperDome? It's because FSB blows chunks under those conditions.

Re:No Surpises Here

[Anonymous Coward]

Pretty much any benchmark will put AMD close to or better than Intel. Clock speed is nearly irrelevent, stop buying into Intel's fud.

Re:No Surpises Here

[ThisIsFred]

After looking at those benchmark results, I'd say: Yeah, sure.

Re:Suck...

[Bun]

And I would *LOVE* it if Intel would support AMD in their motherboards, like they did in the old days. 'course, with AMD's reliance on Hypertransport, that might not be such a good idea...

AFAIK, Intel has never made a compatible CPU for an AMD-develped platform. It was AMD who built their CPUs to be compatible with Intel motherboards/sockets/chipsets. When Intel made the PIV, they either refused to license AMD to build compatible CPUs, or AMD declined to manufacture them.

FUD, yeah...

[Anonymous Coward]

But the Xeon you show isn't the one in the review! The review was for a just released mega-super-expensive 4MB cache version of the 3.4GHZ Xeon.

Parent is absolutely wrong.

[Anonymous Coward]

The parent is wrong, they're not testing CPUs. Testing completely different architectures and believing that 'because we use the same memory the test is balanced' is just too absurd!

Re:Parent is absolutely wrong.

[brucmack]

...so would it be ok then if they used completely different disk arrays? Because, you know, some AMD boards might support RAID differently than Intel, and it's not like the test is balanced anyway, so let's just pick our peripherals out of a hat...

Seriously though, as someone who has had to recommend server purchases in the past, a benchmark like this is much more useful than a benchmark that starts using different parameters. Also consider that server memory isn't cheap, and DDR333 is likely to be in widespread use already. This really improves the ability of interested parties to apply the results to their own systems.

It seems that many disagree, but the main point of my post was that Opteron already wins this benchmark and Anand says so, so I refuse to believe that the choice of memory was meant to unfairly benefit the Xeon solution.

Some?

[Anonymous Coward]

In fact, in some cases, the quad Opteron outperformed the quad Xeon.

Wasn't it in fact in _all_ cases?

Re:Why benchmark games?

[Zeriel]

*raises hand* I'm a corporate IT type, and I read his benchmarks. Along with about three others on a regular basis. Because sometimes, "real work" tends to scale pretty similarly to game performance--especially when that real work is a lot of 3D graphic operations.

Re:Suck...

[Duty]

I'll rephrase that for you: Why can't *anyone* (except Apple and Sun, if they count) sell a dual-processor solution to desktop users?

I imagine VIA's chips would be prime for this.

Order Entry Stress Test

[DarkHelmet]

Here's a part that I can't help but laugh at:

In our infinite desire to please everyone we worked very closely with a company that could provide us with a truly Enterprise Class SQL stress application. We cannot reveal the identity of the Corporation that provided us with the application because of non-disclosure agreements in place.

Okay... So we know what kind of hardware they're testing against, but not knowing what kind of software they're benchmarking? "We're using an enterprise scenario" isn't good enough.

It's nice to look at pretty charts and all, but I imagine anyone who is going to investigate enterprise level solutions is going to want to know EXACTLY what this is being benchmarked on.

Even though I typically tend to trust Anandtech's outlook on things, I'm still kind of so-so on this review. Their forum test is not really externally reproducible and their enterprise test is too vague. I doubt any IT person would weigh this review too heavily when making a decision.

Then again, I could be wrong.

Story submitter -5 Ray Charles

[Loki_1929]

"although dual Xeon configurations almost always beat dual Opterons."

Perhaps the submitter's screen reader doesn't work well with flash, but in the 2-way benchmarks, Opteron was on top twice, and Xeon was on top 3 times. All the 2-way benchmarks were fairly close (within 5%), and the Xeons never beat the Opterons by a margin greater than 1.7%. I don't quite know where 40% wins translate into "almost always" loses. In other words, the story submitter is a moron, or simply didn't look at the article.

"Results were a little varied as 4-way Opteron systems seemed to fare the best,"

Seemed? Let's see, out of five 4-way benchmarks, Opteron won... all of them - performing about 10% better than Xeons each time.

Since when did we start letting Tom Pabst submit articles to /. ?

Note to editors: When the submission is non-sequitur, either reject it or edit it.

Re:Story submitter -5 Ray Charles

[haggar]

Note to editors: When the submission is non-sequitur, either reject it or edit it.

You must be kidding. Do you seriously think that Slashdot editors have more clue than the average submitter?

Re:Story submitter -5 Ray Charles

[Anonymous Coward]

Since when did we start letting Tom Pabst submit articles to /. ?

Since this type of controversy was recognized as a way to generate additional page hits.

Not Surprising

[RAMMS EIN]

``Results were a little varied as 4-way Opteron systems seemed to fare the best, although dual Xeon configurations almost always beat dual Opterons.''

Varied, perhaps, but not surprising. AMD has integrated the memory controller on the CPU, which could explain their getting better when the number of CPUs increases (the Intels being held back by having to go through the same memory controller).

As for Intel winning out on the dual CPU systems, well, they are ahead of AMD in the CPU speed race, aren't they?

Dual Xeon's beat Dual Opteron's? I think not.

[prisoner-of-enigma]

Opteron systems seemed to fare the best, although dual Xeon configurations almost always beat dual Opterons.

Perhaps the benchmarks show the 2P Xeon's doing OK against 2P Opteron's, but for the price of two Xeon MP chips you can buy five Opteron 848's. Rounding that down, I wonder how well the 2P Xeon does against the 4P Opteron? Oops, Anand already though of that. He says "it would not be pretty." Indeed.

Why are there so many Opteron/Athlon64 sockets?

[emil]

AMD made a bundle by keeping a stable SlotA/SocketA while Intel constantly changed the socket.

I realize that Hypertransport changes things, but why can't AMD settle down into a standard Athlon64 socket? Do they know what this will cost them?

Re:Why are there so many Opteron/Athlon64 sockets?

[Hoser McMoose]

They WILL standardize on a socket, it's just that the socket will be Socket 939 and not the current one.

It's pretty much the same story with SlotA/SocketA. They had an initial design that was quickly replaced. The second socket then stuck it out for the duration.

Intel did pretty much the same thing with their P4, initially releasing it on socket 423 and then quickly moving to socket 478 which has lasted for several years now (though it too will soon be replaced).

Markets change, technology changes, and sometimes sockets need to change with them. Remeber that the specification for Socket 754 and Socket 940 for current Athlon64 chips was set in stone about 3 years ago (before the first beta chips tapped out), and a lot has changed since then. AMD has gone to great lengths to minimize socket changes, but there's only so much that they can do.

Re:Why are there so many Opteron/Athlon64 sockets?

[Anonymous Coward]

> AMD made a bundle by keeping a stable SlotA/SocketA while Intel constantly changed the socket.

I rather doubt it. While a small number of tweakers upgrade their CPUs, 99% of AMDs ship out to end users who don't give a crap. Stable sockets make it *slightly* easier for OEM customers.

Intel changed their socket, but they also went from a 133Mhz bus to a 800Mhz one over time. I kinda doubt you'd want to put a 3.2Ghz P4 on a oldstyle SDRAM board.

Re:Suck...

[Loki_1929]

"When Intel made the PIV, they either refused to license AMD to build compatible CPUs, or AMD declined to manufacture them."

Actually, AMD came out with an entirely new "socket" with the original K7, the Athlon. It actually debuted in a slot form factor, but the resulty remains the same - AMD has been independent of Intel sockets since '99. AMD and Intel have numerous cross-licensing agreements, which is why Intel is offering an AMD64-compatible bunch of CPUs fairly soon. AMD could have used socket 423 or socket 478, but why would you want to redesign your chip to work with a socket that'll make it slower? It also would have killed upgrade routes for AMD's existing customers had they gone to Socket 423 when the P4 came out.

Most people seem to forget that the Athlon family was originally competing against, and beating, the P3 long before the P4 arrived. Thus, AMD came up with a chip that has competed against two major products from Intel for about 5 years now. That's impressive no matter which way you slice it.

Same for the PPC 970 ??

[for_usenet]

I may be off-topic here, but is the CPU-memory configuration used by the AMD CPUs similar to that used by the PPC 970 (a.k.a. Apple's G5)? I haven't read a whole lot of technical details on the 970, but it did sound like it. Can anyone verify/confirm ?

Re:Dual Xeon's beat Dual Opteron's? I think not.

[NerveGas]

Yes, the tests weren't exactly apples-to-apples - the outcomes are actually much better for AMD than the graphs would initially appear.

The graphs mean that Opterons with a "measly" 1 meg of cache are beating out Xeons that have (a) four times the cache, (b) 50% higher clock speed, and (c) a price tag that's three times higher.

Hats off to AMD. In times past (K2/K3), price was the only thing they had better than Intel. Now they've got both price and performance.

steve

Opeteron memory bandwidth.

[ebiederm]

The current opteron memory bandwidth per memory controller is 6400MB/s. So it is twice that in a 2 way system, and four times that in a 4 way system. At least if properly configured. The memory bandwidth in an Itanium2 system with it's shared FSB is 6400MB/s total. Opteron should do much better in 2-way and 4-way benchmarks once OS's begin to optimize for a NUMA architecture. But even with out that the performance is quite good. The only thing I have seen that the Itanium2 has is a larger physical address space support. Which is great if you are building a 512 CPU box and irrelevant everywhere else.

Re:Opeteron memory bandwidth.

[ysachlandil]

This is true... if you use a good NUMA aware OS. Otherwise the application/data you are trying to run/crunch on four processors resides in the memory of one processor, and the other three need to bug it all the time to get at it (or worse invalidate each other's caches when writing to it).

So if you don't use a NUMA aware OS, the memory bandwidth could sink below that of a single CPU system.

--Blerik

Oh come on

[Anonymous Coward]

The cost of the motherboard is relatively small; if you can reuse memory and your other system components, you're getting what... about $150 worth of "shaft", and then only if you ignore the fact that you're actually *USING* your machine right now.

I'm trying to say in a nice way that I think you're all wet with this analysis.

don't like going through banner ads/10 pages?

[LoganEkz]

Try the printable version of the article...

Re:Pffft. These Intel vs. AMD flamewars are pointl

[Shadow99_1]

"Currently Intel's latest 3.0+ GHz offerings are spanking Athlon 64s in benchmarks with 32 bit applications."

What a bunch of crap! That's almost as big a lie as Intel makes of AMD cpu's. I didn't even bother reading the rest when you are obviously delusional.

BTW A64 chips can be had for the same prices as their A32 counterparts in the same speed rating, Motherbaords are almost equally expensive whether A32 or A64, & outperform them by up to 30%. ...

Ok well now I have read the rest of your comments & I have to reply to those as well:

"they're going to go to Intel because it has more upgrade possibilities"

What possibilities? They force upgrades way more often than AMD, & are known for being the best money sink for performance users.

"is cheaper than the Athlon 64 for the same level of computing power"

Uh not really, A64 costs the same or less than Intel for comparable performance as long as you don't follow the rule of 'Mhz/Ghz equals performance'. Take a look at pricewatch or Newegg and see for yourself...

"currently performs better"

Hogwash.

"So this is more of a plea for AMD to extend the Athlon "32" line a bit further. Please AMD, don't prematurely kill off 32-bit Athlon chip development!"

Uh they are building them at least through the end of this year, & probably next to some degree. The thing is their is no real future for A32, performance has been decreasing performance-wise in comparison & wasn't keeping up. AMD realized their best bet was to focus on A64 with it's integrated memory controller & higher IPC than A32 has. A64 is a real contender where as A32 wasn't keeping up & they knew it.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[Ian Wolf]

I think time will show (barring any major gaffes) that AMD made the right move. First and foremost, the Athlon XP appeared to be reaching the end of its run. Cranking up the speed of the processor didn't really seem to yield comparable performance gains. In the early days the performance ratings on the XP line were a little iffy, but they were close enough that most people didn't really care. With the XP 3000, and 3200 though those ratings were dubious at best. The speed ratings ratings for the Athlon 64, however, are pretty accurate, even understated in some areas.

I recently upgraded my principle system and at the time, I was determined that the P4 2.8HT CPU was going to be the chip the machine was built around. It clearly trounced the Athlon XP 2800 in all tests for about $100 more. I was about to order when the AMD Atlhon 64 3000+ was released. It outperformed the Intel chip in most areas, was 64 bit, backwards compatible, and only $20 more. In my opinion, and Anandtech and Tom's Hardware agree, the AMD Athlon 64 3000+ is the best bang for the buck CPU out there.

Pricewatch's Lowest Prices are...

AMD Athlon 64 3000+ - $205
AMD Athlon 64 3200+ - $251
AMD Athlon 64 3400+ - $401

Intel Pentium 4 3.0 (800) - $209
Intel Pentium 4 3.2 (800) - $273
Intel Pentium 4 3.4 (800) - $420

AMD Athlon XP 3000+ - $158
AMD Athlon XP 3200+ - $194

Looking at these benchmarks here and on the following page here. You can see, with the exception of pure media encoding, the Athlon 64's perform better. The margin is slim in some areas, and quite large in others. All this and the chips are a little cheaper.

In my opinion, the XP line was dead. It had gone as far as it was going to go. I think AMD knew that the 3000 and 3200 were more like the 2850 and 2900 and they weren't going to get any more mileage out of the design. I definitely think AMD made the right call putting all their eggs in the Opteron/Athlon 64 basket.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[dpilot]

Problem is, much as we might like 32-bit Athlons, AMD hasn't been doing well at them, financially. AMD is in this to make money, not to make hobbyists happy. In the past they have been able to do both, and when they haven't, they've still been pleasing the gearheads. Maybe they'll be able to again in the future. But for the moment, AMD is attempting to climb upstream into the highly profitable small-to-mid server space with Opteron, and that's where their focus is. I suspect that Athlon-64 will get more attention in due time, and you'll be happy, again. At the moment, we're in the gap.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[CritterNYC]

Currently Intel's latest 3.0+ GHz offerings are spanking Athlon 64s in benchmarks with 32 bit applications.

I wouldn't really say that. Sure, if you want to buy the $1,100 Pentium 4 Extreme Edition. But for normal, real-world users, the Athlon64's price:performance ratio kicks the P4's butt.

When excluding the overpriced P4EE and the AMD FX51:

AMD64 Wins: Business Winstone 2004, Content Creation Winstone 2004, Aquamark FPS, Halo, Gunmetal, Unreal Tournament, Warcraft 3, Quake 3 Arena, Jedi Knight: Jedi Academy, Wolfenstein: Enemy Territory, Quake 3 Arena Source Compile

P4 Wins: SysMark 2004*, Aquamark CPU, DivX Encoding, 3dsmax r5, Lightwave 7.5

*SysMark 2004 is listed once, rather than each individual test. It heavily favors Intel and doesn't reflect other benchmarks or real-word performance.

SOURCE: AnandTech

Re:Pffft. These Intel vs. AMD flamewars are pointl

[CritterNYC]

But yes, I agree with you, AMD cannot neglect the desktop market, unless it makes AMD64 cheap enough that it can put them in all computers (which I think is their inevitable goal). Hell, once eMachines starts stocking them in Computer City, I think they'll have achieved it.

The Mobile Athlon64 3000+-based eMachines M6807 latpop is available at Circuit City and Best Buy (M6805).

The Athlon64 3200+-based Compaq s6900NX is also available at Circuit City.

The Athlon64 3200+-based eMachines T6000 is available at Best Buy.

That good enough?

don't forget to control for...

[spamspam]

...power supply, case, cooling solution, installation, etc etc

there are many factors that COULD determine a processor failure - personally i think anybody that tries to throw out the 'i got a bad cpu' argument is a dipstick and probably hasn't built a single system from parts or had more than 1 computer in their lifetime.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[Tassach]

32-bit Athlon isn't going away anytime soon -- hell, (AFIK) AMD is still cranking out the venerable K6-2's.

AMD's appeal has almost always been that they deliver more bang for the buck than Intel. The CPU market tends to be polarized to the 2 extremes: power-hungry users to whom price is no object, and value-concious users on a budget. Yeah, a top-end P4 may spank a top-end Athlon32 (if you consider being a few percentage points faster a spanking), but you wind up paying an absurd premium for that minor performance bump.

Is a small performance improvement worth a huge price difference? Maybe if you have money to burn or if you are using computer speed as a means of compensating for an *ahem* personal inadequacy. Anyone who is interested in getting the most out of his money isn't going to buy top-of-the-line, bleeding-edge chips anyway -- what was top of the line 3 months ago is good enough for most people and is a much better value for the money on a price/performance basis.

Let's say you have a maximum of $170 to spend on a CPU. Looking on NewEgg, you can get a 1.4GHz Opteron, an Athlon-XP 3000+, or a 2.4Ghz P4 without going over budget. Either of these 2 AMD chips is going to spank the similarly-priced Intel chip by a WIDE margin. To get a 3.0GHz P4, you'd spend $219 -- that's a 25% price increase for basically the same performance. If we only have $60 to spend, the difference is even more drastic: a 1.7GHz Celeron w/128K cache versus an Athlon XP 2000+ w/256K cache.

No one likes NUMA?

[Hoser McMoose]

You do seem to be correct that companies don't want to emphasize NUMA too much from the software side of things. AMD has also been known to say that their Opterons don't need any NUMA optimizations.

They actually used the term "SUMO" (complete with a drawing of a sumo wrestler in one of their presentations! :) ), for Sufficiently Uniform Memory Organization.

Regardless of what the marketing depertment says and to who, it's still NUMA. A test on Ace's Hardware a while back showed that a 2P Opteron can see up to a 20% faster with NUMA optimizations, and presumably a 4P Opteron will see an even bigger boost.

I would guess that Anand's test linked in this article had those NUMA optimizations enabled already, though he doesn't specify any BIOS options. He did use Win2K3 which is NUMA aware.

Are there any benchies using Mathematica/Maple?

[Anonymous Coward]

For the platforms mentioned above.

TIA

Misleading summary

[Anonymous Coward]

dual Xeon configurations almost always beat dual Opterons.

Whoever wrote this, and whoever put it on Slashdot's front page, obviously did not read the article. The Xeon wins three of the nine tests, and that's by a 1-2% margin. It loses the rest by 10%.

Re:OH GOODIE, PINGULAR IS POSTING AGAIN! HEY FAG,

[Anonymous Coward]

I don't see Pingular making any insightful comments, either.

re: AMD vs Intel

[Anonymous Coward]

Has anyone noticed how the comparisons of Intel vs AMD always show AMD slightly less than Intel? Has anyone ever suspected that AMD might be faking that it runs at a slower clock speed, with less cache just to get some people saying that AMD "whoops" intel's ass? Theres something not right about an 800 pound gorilla getting beat up by a monkey

God's own plugin: adblock

[akb]

Flash ads got you down? Sounds like you need the Adblock plugin for Mozilla. Allows you to block flash, images, javascript, iframes, etc with regular expressions. I used privoxy before but having to edit text files everytime I wanted to block something sucked, adblock is integrated into the browser so it only takes a click or two.

opteron whipped its ass

[SQLz]

In the two way tests when the Xeon won it was always by a margin of like 1% to 3% where as when the Xeon lost, it was usually by 5-11+%. Not sure how the Xeon fared well here since it has lots more clock speed and lots more cache. It basically got owned hard core.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[Anonymous Coward]

Dude- calm down. Its not like he insulted your mother or anything. The sun does not rise and set on your perception of AMD, and somebody disagreeing with you about a computer architecture is nothing to get worked up about. Go back to suckling on Torvalds teat until you can calm your pear shaped ass down.

HA HA!

[Anonymous Coward]

That's funny. You said "resale value" when referring to PC computer equipment.

Sound cards

[sLaSh_N_bUrN_(.Y.)]

The problem I have with integrated sound is when I am playing mp3s and open a large file(avi, game, ..) my sound cuts out for a second. My last mobo had the same problem until I put a $15 SB in. I thought the AC97 would be ok. I was wrong.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[Loki_1929]

"They only produced two marginal FSB400 processors with the "32-bit" Barton core, and then focused all their attention on the Athlon 64s."

Out with the old, in with the new; sounds good to me.

"People who've made a choice in the past year to go with an AMD-compatible FSB400 mainboard are getting the shaft"

Those who made the choice to purchase a CPU which is at the upper end of an architecture's limits have only themselves to blame. If AMD spent another 20 years trying to make K7 faster, you'd be posting to slashdot in 2024 complaining about how people who just bought the latest Athlons are 'getting the shaft'. What about people who bought 233MHz P1s? What about people who bought 600MHz Slot 1s? What about people who bought 1GHz P3s? Are all these people 'getting the shaft'? Or is this simply an inevitable event in the course of chip development? Look, socket A has been around since S370. Since that time, Intel has gone through... what, 3, 4 sockets? AMD makes CPUs that go from competing with 600MHz P3s to competing with 3GHz P4s, and you continue to complain when they finally reach a ceiling they can't break through.

Look, most experts were looking at the death of the K7 at about 2GHz. They were looking at the architecture, and it simply doesn't do well at a whole lot above that. Yes, some chips will make it to higher speeds with excellent cooling, but those are the exceptions - not the rule. It's a credit to AMD that the massive core improvements from Thunderbird to Barton have kept the K7 in competition for this long. Now that the Mhz train has run out of steam and they can't squeeze any more performance out of K7, there's not a whole lot they can do with it. The extra cache, the FSB jumps - they're just not sustaining K7 performance improvements anymore. The concept of diminishing returns really comes into play at this point.

"AMD is unwittingly forcing them to move to Intel during their next upgrade."

I'm not quite sure I understand this part. Let's see, I can buy Socket 478 board with a 3.2GHz P4, which will be worthless if I want to get to anything above 3.4GHz. Or, I can wait for LGA775, which might be ready some time this summer. Or, I can go to Athlon64's S754, which will hit 4000+ at a minimum. Or I can wait for Socket 939, which should be out some time towards the end of this month, and go even further with the Athlon64/FX lines.

"Currently Intel's latest 3.0+ GHz offerings are spanking Athlon 64s in benchmarks with 32 bit applications."

Pull yourself away from Tomspropagandamachine.com and look at Anandtech, Ace's hardware, or just about anywhere else on the face of the planet. The only 32-bit apps that the P4 wins at all are encoding/streaming benchmarks. When you look at games, office, rar'ing, etc, the A64s put the P4 down like an old dog.

"When users decide to do the next upgrade, they're going to say "hey, I have to replace my mainboard anyway", and they're going to go to Intel because it has more upgrade possibilities, is cheaper than the Athlon 64 for the same level of computing power, and currently performs better."

You've stumbled into the SCOiverse, I think. Are you going to try and tell me that the P4 is cheaper than an Athlon64 3000+ system?? Let's see: P4 3GHz upgrades to ... 3.4GHz max. Athlon64 3000+ upgrades to, at least 4000+, probably closer to 4400+. As for current performance, you've got to be looking at Tom's to even be able to imagine that one. It's amazing what sorts of results you can conjure up when you rig benchmarks by kneecapping the "competition's" (competition? isn't this supposed to be an unbiased review?) products by tweaking driver versions/settings/etc.

"So this is more of a plea for AMD to extend the Athlon "32" line a bit further. Please AMD, don't prematurely kill off 32-bit Athlon chip development!"

Your wish is Jerry Sanders' command! AMD has already stated that socket A will live throughout '04, an

Sigh

[eddy]

I didn't mean to suggest that it's "The Big Thing". I meant it like "the big change in the new revision"

Re:Pffft. These Intel vs. AMD flamewars are pointl

[Anonymous Coward]

I have a hard time giving anyone who cannot type "and" any credibility at all. This is like those who type "ne1" for "anyone". When you can converse like an intelligent person, maybe someone will listen to you.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[ThisIsFred]

Okay, you must have been the guy that moderated me as a troll (lol). Sorry, didn't mean to push your buttons.

The point that I think some of you folks might have skipped over is the upgrade path. This knocks about $150 (average) off the price of the Intel upgrade path, because you can move to their faster P4 CPUs without having to change socket types. The new Prescott CPUs don't require a mainboard with a new socket (assuming your Intel-compatible board is fairly recent). Think about that: If I want to go from my Athlon XP to an Athlon 64, I have to spend another $150 and tear apart my machine. To the average consumer, this means buying a whole new PC. If I had got a socket 478 board, I would just buy the next fastest Intel chip (or whichever was the best balance or price and performance).

A new Athlon 64 may be comparable to last year's P4 chips, but the newest Intel chips with larger on-chip cache are clearly out in front of anything AMD has to offer. I *am* talking about the most expensive chips, but the current price is not relevant, because my whole point was thinking about future upgrades.

Relevant information
Other relevant information

I'm not an Intel marketing drone, I've provided benchmarks to back up what I said, and I think offered some pretty clear reasoning in the process. Come on guys, don't label me a troublemaker. I'm an AMD guy, not an Intel guy. I just don't want to see AMD fail, because that's where I've put my money (I've owned 6 machines with AMD processors so far, currently three of those are in use).

Re:Pffft. These Intel vs. AMD flamewars are pointl

[Anonymous Coward]

wow...that was cutting.

Using a recognised symbol for a word instead of a word....definitely a sign of a lack of intelligence.

using & is definitely not the same thing as ne1 or u r.

Re:Pffft. You are pointless.

[cbreaker]

Intel doesn't have any more upgrade choices then AMD, they never have.

Please state an example, after the "Pentium Overdrive" processors for 486 boards, where can you upgrade your system to a new CPU core without replacing the motherboard and most likely RAM too. I'd like to know - because there isn't any.

The Athlon64 is a direct upgrade for your 32-bit Athlon. Forget about the x86-64 extensions. It's a faster processor even without them. You wouldn't be spewing the garbage if it didn't have this capability.. so why bitch about something that may help you in the future?

Re:Suck...

[Bun]

Actually, AMD came out with an entirely new "socket" with the original K7, the Athlon. It actually debuted in a slot form factor, but the resulty remains the same - AMD has been independent of Intel sockets since '99. AMD and Intel have numerous cross-licensing agreements, which is why Intel is offering an AMD64-compatible bunch of CPUs fairly soon. AMD could have used socket 423 or socket 478, but why would you want to redesign your chip to work with a socket that'll make it slower? It also would have killed upgrade routes for AMD's existing customers had they gone to Socket 423 when the P4 came out.

I'm not disputing that AMD came out with their own 'socket' with the Athlon. I was wrong, in fact, about the generation of CPU that forced AMD's hand.

AMD came out with their own packaging starting with the Slot-A Athlon. This is because Intel refused to grant them licenses to manufacture Slot-1, and then Socket-370 compatible CPUs. As a result of this, AMD was forced to go its own way and develop independent CPU/Chipset x86 platforms. It was either that, or continue making Socket-7 compatible CPUs and spiral down the obsolete technology commode.

I'm pretty sure that, had they been granted a license, AMD would have manufactured Slot-1 and then Socket 370 compatible CPUs and saved themselves the costs of developing their own chipsets, partnering with other chipset manufaturers to develop Athlon chipsets, convincing motherboard manufacturers to build on their platforms, etc. As it was, once they had gone down the Athlon path, there was no going back.

Can Anyone Say 64-bit OS.

[rich115]

...and then they wasted everyones time by running Windows in the benchmark. Why not a 64 bit OS on the Opteron? Linux or Solaris x86 for instance. I'd prefer to see the difference then.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[Grant_Watson]

The new Prescott CPUs don't require a mainboard with a new socket (assuming your Intel-compatible board is fairly recent). Think about that: If I want to go from my Athlon XP to an Athlon 64, I have to spend another $150 and tear apart my machine. To the average consumer, this means buying a whole new PC.

The average consumer doesn't upgrade his/her CPU. To the average consumer, doing so would probably mean buying a whole PC, whether or not a new motherboard were required.

It's us geeks who care.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[juhaz]

You can't moderate and post in the same thread (maybe even story), so clearly he wasn't.

Since when tomshardware and "relevant" fit in same sentence? Besides, future upgrades or no, chips with gigantic cache will _ALWAYS_ be very, very expensive.

Re: AMD vs Intel

[silverfuck]

And how, exactly, might that be accomplished? Remember, these are not results from AMD, these are independant (debate that elsewhere!) hardware labs and tech sites. They set the clock speeds, they read how much cache is accessible to the core, etc, etc.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[ThisIsFred]

Oh yeah, duh. Well, he got his friends to gang up on me. Yeah, that's it!

Re: big cache, big expense. Not necessarily. Both Intel and AMD are producing chips with 1MB L2 cache. They're quite affordable at around $200. Certainly a jump up from no L2 cache on-chip.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[ThisIsFred]

Intel can afford the millions to sink into Itanic because they do things like release filler products in one product line while perfecting the next. It's a common practice: Win98SE/XP Reloaded for example. Apparently you folks are telling me that the 3000+/400 and 3200+/400 was the filler chip.

Are you going to try and tell me that the P4 is cheaper than an Athlon64 3000+ system?? Let's see: P4 3GHz upgrades to ... 3.4GHz max.

Well, people are going to have things like P4 2.4 or 2.66 machines. I didn't realize there was a 3.4GHz ceiling. Why would there be?

SUCK MY ANUS YOU FAGGOT WANNABE TROLL BITCH

[Anonymous Coward]

cAN yOU gET aNY mORE gAYER?

Re:Pffft. These Intel vs. AMD flamewars are pointl

[Loki_1929]

"Apparently you folks are telling me that the 3000+/400 and 3200+/400 was the filler chip."

I'm telling you that the 3000+/3200+ products were used to compete with the higher-end P4 CPUs, and that they represent the upper end of the potential for the K7 line.

" I didn't realize there was a 3.4GHz ceiling. Why would there be?"

Because Intel is having major problems with the Prescott P4s on S478. Supposedly, LGA775 solves many of those problems. Intel is preparing the dump Socket 478 some time in Q2. The fact that Intel has said that the P4 will hit about 4GHz by the end of this year, coupled with the fact that it's currently stuck at 3.2GHz (with 3.4 coming some time later in Q1), means that 3.4GHz is probably the last Socket478 CPU.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[Shadow99_1]

Actually I didn't moderate you (which should be self-obvious since I can't moderate and post in the same article), nor actually would I have done so. I'd rather respond to each thing seperately...

Oh and on that topic:

"This knocks about $150 (average) off the price of the Intel upgrade path, because you can move to their faster P4 CPUs without having to change socket types"

Sure, but that shiny old P4E/P4EE supporting board would also need to be replaced if they wanted a Intel chip beyond this generation. It's just that Intel had a generation change recently and now will be steady again for awhile. Someone who already had a AMD system would need a new board anyways.

Also once AMD switches to 939 & 940 as the only socket types, those will stay the same for ages. Does anyone remember how long Socket A has lasted? Now Socket A did need improvements to things like the FSB on the socket A platform, but with A64 this won't happen since their is no FSB to increase just the data rate of the hypertransport connections...

"To the average consumer, this means buying a whole new PC."

The average user has been taught to upgrade their whole machine anyways. I don't know a single 'average joe' who would even try to upgrade a CPU. They just know when it's time buy a new box because they feel their machine is now to slow. Most OEM's foster this impression (Dell, HP, Emcahines, Gateway, Sony, etc): 'System to slow? Time to buy a new box!'.

"I would just buy the next fastest Intel chip"

Uh from all reports the next generation will need a new board, in traditional Intel fashion. Only one or two iterations of any chip work with a certain series of baords.

"A new Athlon 64 may be comparable to last year's P4 chips, but the newest Intel chips with larger on-chip cache are clearly out in front of anything AMD has to offer."

You listed Tom 'Intel's Bitch' Pabst's site for your info. I wouldn't listen to Tom if you paid me & if you are listening to Tom then I know why your heads so messed up. I've talked directly to Tom, so I'm not talking out my ass either. Him & Anand just need to go into a field they know something about, cause it's not PC's or PC hardware... Find soem real sources who don't go out of their way to put down AMD & I'd lsiten, but even Tech TV (of all sources! Almost as bad as Tom & Anand) has admitted the A64's are the king of the hill of CPU's.

Unfortunatly I can't give you a bunch of links right now as I just redid my whole system, but try Ace's, tech report, digitimes, CPU review, etc... They have a better understanding of what they are talking about & you see A64 win more often than not, and loose only in SSE2 (aka Intel manipulated) or heavily Intel optimized apps (mostly multimedia encoding which Intel invests the most money in).

Re:Pffft. These Intel vs. AMD flamewars are pointl

[Shadow99_1]

I don't know anyone else who posts on slashdot (though I do know a few who read it), so no I didn't.

Big cache is more expensive, because it increases the size of the CPU's die. In fact cache has more to do with chip die size than almost anything. Also cache eats power, so big cache CPU's are: hotter & bigger. Oh and lets not forget that cache is the most likely part to fail on cpu's. Hence why several times in CPU history the low grade of a CPU has been higher end CPU's with half their cache disabled because it was bad.

All of that adds up to them being expensive. I have heard thougbh that Opteron gains very little from more L2 cache & that is the reason why they may lower the cache to 512k from 1M in their A64 line.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[ckaminski]

Yes, but AFAIK, none of those stores actually *HAVE* any.

Could be wrong though. I only checked last week... :-)

I sure hope microsoft has an easy upgrade path from XP32 to XP64.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[CritterNYC]

Yes, but AFAIK, none of those stores actually *HAVE* any.

Actually, the Circuit City stores in NYC I checked have the AMD64 Compaq desktop in stock. And Best Buy in NJ has the AMD64 eMachines desktop in stock.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[nial-in-a-box]

You didn't get the shaft if you bought a FSB400 motherboard. You shafted yourself. Don't think that a CPU that is not on the market ever will be. Buying a motherboard as some sort of future investment is about the worst idea I have ever heard. Yea, sometimes you want the latest and greatest, and it happens to work with your current CPU, but you can't get pissed when you can't the CPU you want for the motherboard later on. This is similar to buying a new truck, minus the engine, and throwing in this engine you already have that happens to work. You have no guarantee that the next engine design will work in your current truck. There is usually no reason to suspect it will either. I am sorry this didn't work out for you. I am disappointed as well, but no one got the shaft here.

I am also perfectly content with the performance of my FSB333 system and I can't imagine why anyone (especially a desktop user) would need anything faster. Honestly, there is no appreciable or noticeable difference in performance between these two FSB speeds. I don't have any benchmarks to back myself up, so you probably should ignore what I'm saying. Realistically, all you're going to see is maybe a handful more FPS in your favorite game, or a couple seconds less on a minute-plus operation. I realize that our demand for the best possible is a good part of what drives this industry, but it also drives this wonderful planned obsolesence which is wasteful and foolish. I'm typing this on a Dell laptop that was top of the line 2 years ago. It isn't anymore, but it still is snappy as all hell in Linux and Windows. It only has a Geforce4go, but it plays almost every game just fine. And you know what, it's only a P4M 1.8 running on a FSB266.

I understand that we always want the best we can get, but we need to understand also that when we're on the cutting edge of technology, we're going to lose some blood. That's just how it works.

Re:Suck...

[Loki_1929]

"This is because Intel refused to grant them licenses to manufacture Slot-1, and then Socket-370 compatible CPUs."

Cite sources, please? I've never heard this before. Ever.

"It was either that, or continue making Socket-7 compatible CPUs and spiral down the obsolete technology commode."

The "Super" Socket 7 may have been aging, but the K6-3 was still a kickass chip that used Socket7 to its full potential.

"I'm pretty sure that, had they been granted a license, AMD would have manufactured Slot-1 and then Socket 370 compatible CPUs and saved themselves the costs of developing their own chipsets"

I seriously doubt that, for a number of reasons. First of all, it had always been Jerry Sanders' dream to one day split off completely from Intel, competing directly against it with superior technology, and eventually surpassing it in performance and market share. Having gotten many board manufacturers behind AMD by extending the life of Socket 7, AMD was able to use the designs they'd been working with for a while, and put them to practical use. What Slot 1 and Socket 370 would have lacked, more than anything, is the Alpha 21164/21264 Bus. AMD had been designing much of the K7 architecture around it, seeing it as a very advanced (wasn't everything when it came to Alpha?) design with very promising potential. The result of this choice is that they were able to stick with Socket A for several years, and continue producing it even today. I think AMD looked at what Intel was doing and saw nothing but performance ceilings. I think they then looked at what Alpha was doing and saw the sky as the limit.

P6 was designed from the ground up by folks who had never done an x86 CPU before. This new team had no x86 experience, and thus presented AMD with a perfect opportunity to split. The timing could not have been better for AMD to throw what weight it had behind the new K7 CPUs. Now, that's not to say that P6 was poorly designed, nor that there was anything especially wrong with the design of Slot 1/Socket 370; just that the timing was perfect for AMD to stick their own flag in the dirt for a change. They took the risk, and it has paid off handsomely.

Re:Suck...

[Bun]

"This is because Intel refused to grant them licenses to manufacture Slot-1, and then Socket-370 compatible CPUs."

Cite sources, please? I've never heard this before. Ever.

It was common commentary in the on-line press at the time. I found this bit here:

"But the most controversial debate was over Intel's apparent decision to lock out its rivals AMD and Cyrix from following in its footsteps, as AMD and Cyrix had done in creating the K5 and 6x86 chips to compete with the Pentium. The slot 1 interface designed for the Pentium II was patented by Intel, all but assuring that the other two companies would not use it for their new CPUs. This has led to a firestorm of criticism from PC users concerned about competition, upgradability and other issues. With AMD and Cyrix deciding to stick with Socket 7, the market has basically "split" here, and we will have to see what happens in the future."

This take was repeated all over the place. I'm surprised you never heard it.

The "Super" Socket 7 may have been aging, but the K6-3 was still a kickass chip that used Socket7 to its full potential.

True...and irrelevant. While with the K6-3 AMD did manage to push Socket 7 to new performance heights, it doesn't negate the fact that that platform was a technological dead end. Every platform is a technological dead end, given time.... so they still would have had a choice to make for the future: piggyback Intel's platform, or produce their own. Since Intel had locked them out, the choice was made for them.

Your take on the rest seems plausible. It's possible that being locked out of Slot 1 was just the push that Jerry Sanders needed to convince the board to go with his dream. At any rate, I for one am glad they did go this route. The K7 and K8 are terrific examples of what can be done with the x86 platform, and have provided the public with an excellent (some would say better) alternative to Intel's offerings - at substatially lower cost.

Re:Pffft. These Intel vs. AMD flamewars are pointl

[ckaminski]

And now they're owned by gateway... all the great work eMachines went through to shed their shitty past... down the tubes... :-/

Thanks for the info!

-Chris