Intel Itanium 2 Benchmarks

Pablo writes "Over at VR-Zone we saw some interesting benchmarks of the upcoming Intel Itanium 2 processor codenamed McKinley that is on schedule to be launched during second half of this year. With a faster 3MB on-die L3 cache, 6 instructions/cycle and 6.4GB/s of bandwidth, it is poised to perform at 1.5-2x of the current Itanium processor. There is an overview of how the Intel Itanium 2 at 1Ghz clock frequency will perform against the current Itanium 800Mhz and Sun's Ultra Sparc III RISC processor."

Comments

Re:AMD

[robburt]

At least the Itanium is large enough for me to put my coffee cup on to keep it warm. =)

Strange

[morbid]

Isn't it strange how they always compare their processors to competitors' processors of a lower frequency?

Re:Strange

[Anonymous Coward]

Not if the competitor doesn't have anything comparable. Welcome to Marketing 101, now sit down.

Re:Strange

[morbid]

What, like UltraSPARC III at 1050MHz with 8 megs of cache?

No benchmarks

[KingKire64]

This is just a Marketing Piece put out by intel. All the "Benchmarks" are proposed Estimates. And why would a dinky website get a hold of something this "Big"? Dont know just questions.


Mod Me down Please

Re:No benchmarks

[Anonymous Coward]

Can some karmawhore please post a copy since the site is /.-ed

Re:No benchmarks

[Merlin42]

Personally I like how each page in the ppt presentations used a slightly different USIII ranging from 800 to 1050Mhz. Hmm smells like marketing picked out the best from a bunch of (simulated?) benchmarks. Everything was labeled 'simulated' or 'estimated'.

Nothing to see here folks please move along.

Well other than the slow death of competing high end architectures.

Lets see here we have:
SPARC ... still competitive I think
Itanium ... taking over
Alpha ... going the way of the dodo (im really sad about this)
PA-RISC ... transitioning to Itanium
MIPS ... never really liked them for big compute stuff, lets hope SGI can turn things around.
Power4 ... still competitive in performance but AFAIK to get a high end system you need to give your first born to IBM. And, IMO not really designed for the HPC kind of stuff I'm interested in.
Cray ... ? I've heard of some really cool stuff being developed, i'll believe it when I see it.

What eles is out there? I haven't really been in the market for a high end system in a while, but it feels like the market is shrinking and soon Itanium will be "the choice," unless legacy support is a concern .

Re:No benchmarks

[T-Punkt]

Even the graph has been done by a marketing guy.
They sorted the benchmark results in ascending order and the connected the data points of completely different and independet benchmarks by a line!

What shall the line tell you? The faker the benchmark the better the results? Or
"This is a line graph that doesn't make sense at all. But look: It shows an increase, increase is good, so Itanion 2 is good!"

Re:No benchmarks

[pmz]

Absolutely. From one of the slides "All projections based on Intel estimates[emphasis mine]...using...workload testing at Intel[emphasis mine]."

And, absolutely none of the benchmarks are substantiated with real data!!!

Only a fool would accept any of this presentation as fact. An even bigger fool would use this presentation in a decision whether to buy Sun or Intel.

Re:No benchmarks

[Anonymous Coward]

I agree. Much of the data is inconsistent with what I have seen of Sparc III vs Itanium (original).

Late last year, at a high-performance computing conference, at a presentation by a guy from INTEL, he showed some benchmarks of 5 numerical applications on a 500MHz USIII versus an 800MHz Itanium. On one of the 5 apps, the Itanium had nearly twice the performance. But on the other 4, the sparc beat the itanum hands down, factors of two or more. The intel guy attributed this mostly to compiler technology. That is plausible, but not something that can be fixed easily. I see from their 'projected estimates' that they only attribute 5% increase in itanium 2 performance to compiler improvements.

But itanium 2 apparantly beats a USIII 1050MHz in every single benchmark, by a wide margin ??? The numbers just don't add up.

Re:No benchmarks

[Anonymous Coward]

The reason they compare to different Sparc is because not all benchmarks are available for the newest one (Sun is playing the game too and is only publishing benchmarks where it serves marketing).

Note that Sun has been cheating on Spec, that's the only way they can make recent Sparcs look competitive. It's a shame that it will force every other vendor to cheat in the same way.

Re:No benchmarks

[Anonymous Coward]

You can't compare a Cray to all these dinky server processors, you dumbass.

Re:No benchmarks

[T-Punkt]

Cray used (or uses?) SPARCs (CS6400) and Alphas (T3 line) in many of their systems. The T3s are the fastest Crays ever built (see www.top500.org).

Re:No benchmarks

[binaryDigit]

Power4 ... still competitive in performance but AFAIK to get a high end system you need to give your first born to IBM. And, IMO not really designed for the HPC kind of stuff I'm interested in

Hmm, what makes the Power4 not suitable for your HPC needs. IBM seems to think it's more than fine since it's used in ASCI White (and anything else they sell for that purpose).

Re:No benchmarks

[Flywheel]

No INO PA-RISC is not transitioning into Itanium, HP did try that but the performance loss was not acceptable. What I think might happen is that PA-RISC is abruptly killed and Itanium replaces it directly.

MIPS is dead on the workstation/server scene, SGI went the Itanium way...MIPS is today almost only for embedded devices.

SUN is building SMT into a variant of the next US generation, used for Quad machines and bigger.
But remember the strength of a SUN SPARC machine is not the processors, which hasn't been cutting edge for many years, but the big picture (Overall performance) - the machines are extremely well build.

But lately I've heard rumours about Compaq reconsidering the death of the Alpha - because of the possible future Itanium2 flop, even internal forces of the HP part of Compaq is reconsidering the death plans of PA-RISC.

Re:No benchmarks

[joib]

Power4 ... still competitive in performance but AFAIK to get a high end system you need to give your first born to IBM. And, IMO not really designed for the HPC kind of stuff I'm interested in.

Huh??? What are you smoking? :) The HPC market is one of the primary markets for the power4, the other being big enterprise systems. In fact, CSC, the Finnish national supercomputer center is currently installing their new toy, a power4 machine which will have 512 processors when it's finished in september. FYI, that's 256 power4 chips, as one chip has 2 cpu cores. Anyway, the design consists of 16 fairly standard 32 cpu pSeries 690 refrigerator sized boxes. Currently I think they are connected with gigabit ethernet or something like that, but during the summer a proprietary IBM high speed interconnect will be installed. Total performance is estimated to be about 2.2 teraflops, more than 4 times faster than the old 540 cpu Cray T3E, and placing the computer among the fastest in Europe. Currently I think 6 nodes are operational...

BRAG MODE ON
And I have an account on that baby!!! *Drool* Wonder how many fps quake would get? ;-)
BRAG MODE OFF
No seriously, they naturally have a strict policy on what you are allowed to run on it. You have to fill out forms requesting cpu hours with project descriptions etc. etc. Anyway, my plan is to run ab initio calculations on it. Hopefully that is. They're having some serious problems, related to MPI, I think... Which has led to the fact that everyone is submitting the big jobs they planned to run on it to the old T3E, which is rapidly getting overloaded...:(

Re:No benchmarks

[Merlin42]

Ok, guess I'm a little out of the loop. In my rather limited experience I have only ever seen Power4's used for database/mainframe kind of stuff. For example, at a previous job we had an AIX box running the backup/tape robot machine and used Alphas for compute tasks. Most of the marketroid crap i have heard relating to power4 has had to do with reliability at the hardware level. IIRC they have multiple identical cpus that execute the same instruction stream and check the results, or is that only in the BIG mainframes?

Re:No benchmarks

[Merlin42]

I sure hope you are right about Alpha or PA-RISC.

Has SGI really gone Itanium? They have waffled on a LOT of things for years now. They kind of went wintel then backed out, then kind of went x86 linux then backed out. Are they planning any new ia-64 products? The 750 is a legacy product and the Pro64 compiler seems to be gone.

Re:No benchmarks

[Anonymous Coward]

Late last year, at a high-performance computing conference, at a presentation by a guy from INTEL, he showed some benchmarks of 5 numerical applications on a 500MHz USIII versus an 800MHz Itanium.

Even more bogus: the slowest US-III I've seen was 750 MHz. I don't think Sun directly sold anything slower, especially in benchmarkable non-beta systems, and I'm not sure if 500 MHz chips / systems were even made.

On the other hand, US-II family chips do come in 500 MHz speeds.

Re:No benchmarks

[Anonymous Coward]

If Sun improves their compilers in a non-benchmark special way, is that cheating?

Re:No benchmarks

[joib]

At least some older T3s used MIPS processors (probably during the time they were owned by silicon graphics). Nowadays they seem to be concentrating on their new SV2 vector supercomputer, which has cray designed processors. I guess developing a successor to the T3E is not really viable any longer as clusters like the IBM system I descibe in the above post or even el cheapo Linux/Intel clusters do essentially the same thing as a T3E type computer at significantly lower cost. In fact, Cray sells Linux clusters these days (iron provided by Dell).

Re:No benchmarks

[Anonymous Coward]

"MIPS is dead on the workstation/server scene, SGI went the Itanium way...MIPS is today almost only for embedded devices."

I'd love to know where people hear these kinds of things! I'd like to find the source and plug it good.

MIPS is SGI's primary platform for their worstation and server product lines. They will shortly be releasing Itanium based servers running Linux but they have stated again and again that MIPS/IRIX and ITANTIUM/LINUX are seperate product lines. Some of SGI's troubles stem from the fact that Intel is 2+ years late brining Itanium to market, they bet the farm on somone elses vaporware instead of their own (H1 & H2).

Re:No benchmarks

[Anonymous Coward]

It is a great box but only if you use half of the CPU's. not enough cache to go around on the dual cpu boards so IBM has to disable half of them and sell them as 16 ways. Kinda sad.

And their switch architech for connecting them together is slow and has a relative high latency. But for what they are charging...

It is a SUN killer not a HPTC box. They have been telling their customers for the last two years that this box is for TPC not HPTC. HPTC low margin. TPC high margin. They want the HPTC customers to migrate towards linux.

Thier switch is not even really ready for prime time on these boxes.

Re:No benchmarks

[flynn_nrg]

"to power4 has had to do with reliability at the hardware level. IIRC they have multiple identical cpus that execute the same instruction stream and check the results, or is that only in the BIG mainframes?"

Yes, that only applies to the really big iron, i.e. S/390 mainframes. The other boxen use redundant caches a la Sun's SPARC processors. Keep in mind that running N redundant processors is extremely expensive, not only putting the extra CPUs, but the 'magic' hardware required to compare every instruction executed by them.

Re:No benchmarks

[Flywheel]

It was an official statement I fell over a few years back, the plan is/was to migrate to Itanium .... the problem for MIPS is that SGI is almost the only costumer buying these wonderfull processors.
The sales of MIPS processors for embedded devices is quite good....AMD Alchemy is also MIPS based (Current is MIPS 32, the next generation will be MIPS 64).

Re:No benchmarks

[javiercero]

NO! The T3s never used mips EVER! You are thinking of the origins 2K's that SGI relabeled as CRAY.

Re:No benchmarks

[joib]

Uh, yes. You're correct. T3:s are all alpha based. Guess it was just some idea I got when I was on a tour of the local supercomputer center a few years ago (they have a T3E with 375MHz CPU:s)...

Re:No benchmarks

[joib]

It is a great box but only if you use half of the CPU's. not enough cache to go around on the dual cpu boards so IBM has to disable half of them and sell them as 16 ways. Kinda sad.

And their switch architech for connecting them together is slow and has a relative high latency. But for what they are charging...

It is a SUN killer not a HPTC box. They have been telling their customers for the last two years that this box is for TPC not HPTC. HPTC low margin. TPC high margin. They want the HPTC customers to migrate towards linux.

Thier switch is not even really ready for prime time on these boxes.

Here you can see some benchmarks comparing the 32-way p690 with the 16-way HPC (ie. same machine but with only one cpu per power4 chip). The HPC achieves about 80% of the p690 performance. While this is impressive and suggests that the 2-cpu power4 suffers from L2 cache bandwidth I still think this is a semi-low-end configuration. When you want many cpu:s the internode bandwidth and latency is probably going to be more important. I.e. if you want say 512 cpu:s then 16 32-way p690:s are probably going to be faster than 32 16-way HPC:s, depending on your application of course. Not to mention reduced space, electricity and cooling requirements.

And also IIRC their switch so far supports only 16 nodes, although IBM claims it will support 128 nodes by the end of the year. And I wouldn't call it slow either, speed is 1GB/s (that's gigaBYTES, not gigaBITS), which is about the same as the T3E IIRC.

Finally, I'd say it's very much a HPC box in addition to a SUN killer. Take a look at the top500 list. There's lots of IBM SP:s there at the top isn't it? The p690 is essentially a SP with the 375 MHz power3:s replaced by 1+ GHz power4:s. Same goes for the switch. It's an upgraded version of the SP switch.

Of course all these IBM machines are seriously smoked by this new Japanese earth simulator machine, which is supposed to be 5 times faster than the current top500 #1, the ASCI white (an IBM SP with 8192 cpu:s).

Re:No benchmarks

[(outer-limits)]

Hey, you left out IBM mainframe processor, still going strong. And if you check out the PowerPC, it is very strongly based on the mainframe processor design, (simplified, fortunately). The power PC is really just an improvement on a processor that first saw the light of day in the pot smoking, tie died '60s. The processor that IBM now uses in its Z/OS systems is really just a tweaked power pc processor.

Re:No benchmarks

[Anonymous Coward]

and only an even bigger fool doesn't realize
that silicon is already there and performing
99.99% of these numbers already.

chill people.

Re:No benchmarks

[(outer-limits)]

Which is the big problem for the Itanium, make a fast but difficult to program processor, and let the software sort it out.

That sort of thinking had intel producing the Pentium PRO, optimised for running in 32bit mode, slowed down by all the 16 bit software everyone was running. It was years before MS finally brought out NT, a true 64 bit operating system, that was really capable of exploiting the potential of the PRO.

I'll wait for the real product

[Anonymous Coward]

It's all fine and dandy for servers that are loaded up with 8gig+ of ram, but do I really need 16gigs of ram for surfing & email?

Wait a minute, what am I talking about. Sure I need 20gigs of RAM! Bring it on baby, cuz I'm gonna charge it all to bill gates anyway for the super-duper gaming system with 20gigs of ram, 10 terabytes of disk storage, 3 24" lcd and every other geek toy under the sun.

Re:I'll wait for the real product

[Anonymous Coward]

This is the gayest post i've ever read on slashdot, and, since i've read a lot of gay posts on slashdot, that's really saying something.

On schedule???

[JFMulder]

...Itanium 2 processor codenamed McKinley that is on schedule...
Yea, that's if you forget the part about the first itanium being about a billions years late. :-p

Honestly, I am anxious to see what will come out of this war between AMD and Intel for the desktop market. Too bad they didn't have a comparison between McKinley and AMD's SledgeHammer, since they are destined to the same market.

And I would have posted earlier, but I was slowed down by the slashdot effect!!! :-)))

Re:On schedule???

[Anonymous Coward]

SledgeHammer won't compete with McKinley, but with McKinley's successor, Madison.

Oh, good luck to AMD...

It better be quicker

[Anonymous Coward]

This is the first chip that they've unlocked the die-control on. The chips in the last three series not including server models are set to run with restrictions to prevent overheating even with a fan on them. However, because of the new molex composite, there's less of a concern with stress fractures because of improper channeling of the heat conduits. So it'll actually run even faster than the benchmarks let on once out of beta.

Re:It better be quicker

[Anonymous Coward]

*cough* Intel Marketing Employee *cough*

Re:It better be quicker

[Paul Komarek]

"So it'll actually run even faster than the benchmarks let on once out of beta."

Bullshit. You don't know this, and Intel's history with new processor development suggests otherwise. Maybe this means that they finally have the heat issue under control, and can *finally* reach the clockrates they want to advertise. Maybe it means clock distribution is reducing chip performance to the point that heat isn't an issue. Maybe it means that you work for Intel marketing, and think you are repeating something you heard an Intel engineer saying at a party.

-Paul Komarek

Yamhill?

[ultrabot]

But what's with all the stuff regarding MS urging Intel to use AMD's x86-64? Isn't the future of IA-64 rather bleak right now? Even HP apparently says that "market will decide" whether PA-RISC or IA-64 will be their future Unix platform... Which would not be the case if IA-64 was obviously superior.

Well, this can only mean good for Linux...

Re:Yamhill?

[morbid]

M$'s influence is negligible on all but the tiniest (1-4 way 32-bit) servers. So, on larger servers, you're talking about real OSes and huge applications for which 64-bit versions already exist (have for the best part of 10 years). To port to a new 64-bit arch. with modern compilers and libs, it's not much more difficult than a recompile. So itanic may make some inroads into the server space, but only amongst those who buy into the intel brand name and the hype.

Re:Yamhill?

[Anonymous Coward]

No, HP clearly said that IPF (the official name for IA64 now) is going to completely replace PA-RISC / Alpha / Mips in the next 5-10 years.

The market has decided already. It voted for "cheaper", and even HP can't spend the money to keep those other processors competitive without pricing its machines out of the market.

I don't give more than 5 years to Sparc

Re:Yamhill?

[kawaichan]

I think IA-64 should been seen as a long term investment for Intel, it's a good piece of technology (well, at least it was expensive) and it's just not up for prime yet.

x86-64 on the other hand, is a good solution for today's demand.

So who will have an upper hand 5 years from now? probably intel, but who knows, Ghz sells.

Re:Yamhill?

[conway]

Umm, can you plese point to where HP said that?

If HP has been consitent on one front, its that PA-RISC is being phased out and IA-64 phased in. The question is how fast, and how long will the be able to milk the PA-RISC for all the support. (But they're not even betting on that too much it seems, since the IA-64 HP-UX comes with PA-RISC binary emulation -- its really more of runtime translation than emulation -- you get about 80% orignal speed, pretty nifty :) )

Re:Yamhill?

[Zeinfeld]

M$'s influence is negligible on all but the tiniest (1-4 way 32-bit) servers. So, on larger servers, you're talking about real OSes and huge applications for which 64-bit versions already exist (have for the best part of 10 years).

I don't know what you mean by a 'real OS' however having worked at the extreeme high end I can tell you that there are plenty of real fast machines with O/S that are in most respects (except performance) pure junk. And by high end I mean that some of the machines I worked on 10 years ago still outperform top end desktops.

The fact is that high performance machines are usually bought for fairly narrow purposes and as a result tend not to need a lot of an operating system. Back in the early 1990s an awfull lot of computationally intensive work was still being run on IBM mainframes running MVS and JCL which in many respects is not a whole lot better than MSDOS but you would never get the people who ran those piles of junk to admit it.

If you want high performance from a multiprocessor machine the architecture of the O/S does matter a lot, but may not determine the outcome. consider this analogy, when it comes to writing an optimizing compiler modern languages such as Eifel or Java give the compiler writer a heck of a lot more help than a language like Fortran. However until very recently many of the top benchmarks for compiled code were for Fortran compilers simply because brute effort could be used to compensate for poor architecture.

When it comes to the 'architectural features' required to make a multiprocessor machine work fast the cards are all with Microsoft. WNT was designed to work well on multiprocessor platforms and the design team were mainly DEC ex-VMS people who had a lot of experience in that area.

UNIX was originally architected for a UNIprocessor and there are a lot of design decisions that you just would not take if making it easy to run fast on multiprocessors was your objective. On the other hand this matters less than it might because there has been a lot of work since on compensating. The cost being that to make an SMP system work well under 'UNIX' often means having to use features that are proprietary.

The claim that UNIX has a better architecture than WNT is essentially as unprovable as the claim that vi is better than emacs. There are certainly still people in the computing world whose only experience of editing programs is with vi in line mode but will nevertheless post many gigabytes worth of posts to USEnet arguing the point. Having actually worked on O/S design, having used 20 odd O/S and having done system level programming on 4 (including UNIX) I can tell you that UNIX certainly did not get where it is on the strength of design merit alone. Many of the internals of UNIX are as confused and obfuscated as the syntax of the csh.

To port to a new 64-bit arch. with modern compilers and libs, it's not much more difficult than a recompile.

If you have nice 64 bit clean code that may be the case. The problem is that most people don't start from good code and even if they have tried to keep the code base clean they may not have succeeded. But remember that WNT has already been ported to a 64 bit architecture (Alpha) and although WNT is no longer supported on Alpha you have to believe that the compiler rules are still in place to detect code that is not 64 bit clean.

The issue that is probably more important for Microsoft is how .NET performs on itanium. With .NET they have the ability in theory to take applications and compile them for any architecture at installation time. I don't think there can be much doubt that this architecture is designed to support Itanium.

Re:Yamhill?

[morbid]

"When it comes to the 'architectural features' required to make a multiprocessor machine work fast the cards are all with Microsoft. WNT was designed to work well on multiprocessor platforms and the design team were mainly DEC ex-VMS people who had a lot of experience in that area."

So, you've never heard of Solaris, then? Maybe you should study Solaris internals and you would realise that, although WinNT may be better in some respects than many older OSs, Solaris kicks its ass clean out of the arena. Solaris has an extremely sophisticated internal threading model and scales nearly perfectly linearly to over 100 processors. It's lean and clean. Can you say the same of NT? What iron has NT ever run on with more that 16/32 processors? Have you ever seen the performance tests of NT scaling on such hardware? I think the line is, any more that 4 processors and you just bought a load of useless processors.

Solars is _the_ most advanced UNIX.

Yes, I know NT was designed by the VMS guys, but M$ ruined it post NT3.51 by putting a load of crap to appeal to gamers inside the kernel that shouldn't have been there. That sacrificed reliability and portability.

I hear what you say about the IBM mainframe stuff. We used to track our nuclear fuel inventry/fuel cycle on a 360....

Even Linux is catching up in terms of multi-processor scalability.

The point is, NT (2k, XP etc.) still doesn't have the features you want from an enterprise-class OS and it's a good 5 years behind Linux, let alone Solaris.

Re:Yamhill?

[ultrabot]

Umm, can you plese point to where HP said that?

No, I can't. It was said to a colleague, by a HP representative.

Re:Yamhill?

[morbid]

"To port to a new 64-bit arch. with modern compilers and libs, it's not much more difficult than a recompile.

If you have nice 64 bit clean code that may be the case. The problem is that most people don't start from good code and even if they have tried to keep the code base clean they may not have succeeded."

Well, yes, especially if they have been unfortunate enough to have to develop their code for NT. I'm not talking about desktop apps here, I'm referring to man-size ones like Oracle, SAP, etc.

"But remember that WNT has already been ported to a 64 bit architecture (Alpha)..."

Er, yes, but in the 32-bit mode of the Alpha processor, so it was still a 32-bit OS.

"and although WNT is no longer supported on
Alpha you have to believe that the compiler rules are still in place to detect code that is not 64 bit clean. "

Please refer to my point above.
NT never ran 64-bit on Alpha or Mips. These chips had compatibility modes for running legacy 32-bit code, in which NT ran, much the same was as AMD Hammer (Opteron) has a 32-bit compatability mode so you can run your legacy 32-bit OS while you wait for a 64-bit one to come along.

Re:Yamhill?

[dreamchaser]

That sounds very credible...not...Every HP rep I've ever talked to about it says the same thing; PA-RISC will be phased out in favor of IA-64.

Re:Yamhill?

[Zeinfeld]

So, you've never heard of Solaris, then?

Slowlaris, oh yes, I had Sun sell me that when it was still a real heap of junk. Took the sysadmin a day each to configure four machines. He never did manage to get the multimedia (sound, video camera) features to work, nor had anyone else in the building. You might have heard of it, 545 Technology Square Cambs, Mass.

If they can't get an O/S to function two years after release it does not qualify in my book as superb...

Solaris kicks its ass clean out of the arena. Solaris has an extremely sophisticated internal threading model and scales nearly perfectly linearly to over 100 processors.

As you will note in my original post I differentiated the architectural contributions of Ken Thomson and co from later developers who just had the misfortune to start from the constraints of their design. It is possible to get a UNIX shell to run on a machine whose internal architecture bears little resemblance to UNIX.

The question is what the application developer is required to do to take advantage of the result. Does their application work in scalable fashion across all 100 processors if written for 'UNIX'? Or is it the case (of course it is) that to scale the application must either be 100 independent applications running on each processor or if it is genuinely one application does it have to be rewritten for the specific UNIX variant?

In the old days some processors would have a mode switch. So the VAX could emulate a PDP-11 but while it was doing so it wasn't doing the types of thing you would expect of a VAX, like virtual memory or such.

I have written code for machines with several thousand processors. Absent some development in hardware I may be unaware of nobody has a machine that can load more than about 16 processors that does not have some pretty significant limitations. There is a good reason for this, any given generation of silicon can support a certain processor technology and a certain interconnection technology. Unless you go to some form aof non shared memory system performance tends to top out at 8 processors. Yes I can believe that Sun showed off a machine that can calculate the mandelbrot set on 100 processors with perfect linearity, I doubt that it responds in that manner to all applications. Even exotic MIMD machines don't do that when they are purpose built for the application.

What iron has NT ever run on with more that 16/32 processors?

If Intel are right and they can provide 10 times the performance per processor over Sun then Windows need only run on 16 processors to kick Sun's ass.

However since the benchmark in question compared a 2 processor Intel machine with an 8 processor Sun machine the nature of the game should be obvious. Intel clearly chose a problem for which Sun's platform does not respond with near perfect linearity and the benchmark is obviously skewed in that respect. Even so, if Intel have the higher raw processor performance they have the engineers to build machines with more processors. Intel has at various times developed extreemly fast research prototypes (Bill Dally's hypercube being an example. Intel spends 3.2 Bn on processor research a year, Sun spends only 2 Bn on all its research.

Asking the number of processors NT runs on when Intel is only starting its 64 bit processor line is kind of beside the point. The real issue is where we will be in 3 years time...

Re:Yamhill?

[morbid]

"Slowlaris, oh yes, I had Sun sell me that when it was still a real heap of junk. Took the sysadmin a day each to configure four machines. He never did manage to get the multimedia (sound, video camera)
features to work, nor had anyone else in the building."

Get yourself a decent admin.

"You might have heard of it, 545 Technology Square Cambs, Mass."

Nope. Should I?

"Even exotic MIMD machines don't do that when
they are purpose built for the application."

What, like itanic? :-)

"If Intel are right and they can provide 10 times the performance per processor over Sun then Windows need only run on 16 processors to kick Sun's ass."

So, you believe intel's FUD and marketroids. UltraSAPRC isn't the best on sheer processing power but it's a mature 64-bit platform. Intel has managed to hoodwink everyone else into killing off their (working) 64-bit architectures (PA-RISC, Alpha, MIPS) for some vague promise of Jam Tomorrow(TM) with itanic.

As for hypercubes, the ones I saw were Transputer machines. Never wrote any code for them though. They only let me loose on the VAX cluster with 10 lines of FORTRAN *puke*.

"As you will note in my original post I differentiated the architectural contributions of Ken Thomson and co from later developers who just had the misfortune to start from the constraints of their design"

The design is quite old, but, you know, it's all about presenting a POSIX/SUS etc. front end to the world. Internally virtually all moder UNIXes share about 0% in common with the original UNIX.

"It is possible to get a UNIX shell to run on a machine whose internal architecture bears little resemblance to UNIX."

Quite. However, a shell does not an operating system make. A shell is just an application as far as the OS is concerned, and if you're talking about Free and Open SOurce shells, well they've been _designed_ to be portable. You're not surely suggesting that OS/2+bash == UNIX?

Other than that, I fear that you have ben had by Microsoft's FUD too. There is a famous quote : "Those who do not understan UNIX are condemned to reinvent it - poorly." If you study the internal architecture of NT you will find that, despite M$'s continual assertions that UNIX is dead, outmoded etc. etc., they have come up with a kernel that, with every new release, is looking more and more like some of the big commercial UNIX kernels of 5-10 years ago.

Re:Yamhill?

[Zeinfeld]

"You might have heard of it, 545 Technology Square Cambs, Mass." Nope. Should I?

MIT LCS and AI Labs, if they can't get Solaris to work properly without spending a week doing it the problem is the o/s vendor.

We thought this somewhat strange since at the time SunOS had a realy good reputation on campus. Then we discovered that what people were using was a custom distribution cut by MIT...

As for hypercubes, the ones I saw were Transputer machines. Never wrote any code for them though. They only let me loose on the VAX cluster with 10 lines of FORTRAN *puke*.

You can't build a hypercube from a device with 4 links, its a geometric fact that you need 2 links per dimension. I did use Transputers, including some of the 1000+ node machines. Perhaps if you had some experience of the machines you are ignorantly blathering about your posts would not get modded down.

Your entire post confirms the argument I was originally making, that the term 'UNIX' no longer describes the system internals, it merely describes the application interface. However the application interface is still a major constraint on application design.

If you study the internal architecture of NT you will find that, despite M$'s continual assertions that UNIX is dead, outmoded etc. etc., they have come up with a kernel that, with every new release, is looking more and more like some of the big commercial UNIX kernels of 5-10 years ago.

Clearly from the tone of your post you don't like Windows, I have great difficulty therefore in believing that you have spent such a great deal of time examining the WNT internals that you can come up with such an assessment.

It would not be so surprising if WNT shared some of the concepts in Mach given that Rashid has been working for Microsoft for some time. However the architectural principles have from start to finish been based on VMS.

Re:Yamhill?

[morbid]

"MIT LCS and AI Labs, if they can't get Solaris to work properly without spending a week doing it the problem is the o/s vendor. "

I don't know, I've seen some Astrophysics professors who can't use a telephone...

"You can't build a hypercube from a device with 4 links, its a geometric fact that you need 2 links per dimension."

Er, um, ncube used to make some. I think they had some additional hardware.

"I did use Transputers, including some of the 1000+ node machines. Perhaps if you had some experience of the machines you are ignorantly blathering about your posts would not get modded down. "
My posts get modded down because I am not an IBM sycophant like half the trolls on here. Now you really are flame-baiting!

"Clearly from the tone of your post you don't like Windows, I have great difficulty therefore in believing that you have spent such a great deal of time examining the WNT internals that you can come up with such an assessment."

I spent absolutely zero time studying the NT internals, however my lectured did and he gave us quite a comprehensive description of NT, VMS, Solaris, SysV R3 SysV R4 and some of the really old stuff.

So I reckon you're just a very subtle troll. Sorry you had to result to name-calling and I've got wound up and taken you up on it.

I've been bitch-slapped, so I've got nothing to loose. Karma 50 and still posting at zero!
Put that in your pipe and smoke it, lacky boy!

Re:Yamhill?

[morbid]

As for hypercubes, we had an old order-7 one (128 processors) and a new order -5 one(32 processors) which was actually faster. It was shared between the maths,physics and engineering departments.
Then there was something completely different - the Paramid. It was a pyramid configuration and each node contained an i860 (intel's first 64-bit chip) for the fast maths and a T8000 (transputer) for inter-node communication.
But I'm showing my age.

Re:Yamhill?

[Zeinfeld]

I spent absolutely zero time studying the NT internals, however my lectured did and he gave us quite a comprehensive description of NT, VMS, Solaris, SysV R3 SysV R4 and some of the really old stuff.

So you think your second hand account of some unspecified lecturer's opinions on O/S internals gives you the authority to pontificate on the issue?

I have first hand experience of VMS, NT and Unix internals. I suspect that either the lecturer was as opinionated as yourself, or more likely you didn't spend very much time listening to him either.

As for being a troll, if as you claim you are at Karma 50 and post at zero it sounds to me if a member of the slashcrew has come to an uncomplimentary opinion of you.

Re:Yamhill?

[morbid]

"So you think your second hand account of some unspecified lecturer's opinions on O/S internals gives you the authority to pontificate on the issue?"

Well, from what he said, NT was nearly there, but not quite. It does have an advanced threading model.

"I have first hand experience of VMS, NT and Unix internals. I suspect that either the lecturer was as opinionated as yourself, or more likely you didn't spend very much time listening to him either."

Opinionated is my middle name. However, I have to take his word for it. The guy is very experience in such matters and has written a lot of kernel-level networking code.

"As for being a troll, if as you claim you are at Karma 50 and post at zero it sounds to me if a member of the slashcrew has come to an uncomplimentary opinion of you."

Exactly my point. I have an uncomplementary opinion of them too. I suspect that your mouth is also larger than your brain. I have a new account from which I post non-controversial opinions. I'll keep this one for speaking my mind (what's left of it).

This correspondence is now closed.

Re:Yamhill?

[Zeinfeld]

Well, from what he said, NT was nearly there, but not quite. It does have an advanced threading model.

NT does not have the UNIX thread model, or rather it has pthreads but the system apps all use a different model. This is not suprising if you know about thread models. pthreads is not generally considered the acme of good design. It is based on Tony Hoare's monitor design which is almost thirty years old now and Hoare introduced a much better design only a few years after monitors. However MULTICS was designed in the intervening years and when SMP Unix systems started to emerge in the late 80s the work was largely driven by the MULTICS work.

Opinionated is my middle name. However, I have to take his word for it. The guy is very experience in such matters and has written a lot of kernel-level networking code.

I don't know who your lecturer was but I'll bet he wasn't Tony Hoare, my Oxford Tutor who along with Per Brich Hansen and Dijkstra invented most of the concepts on which all modern O/S concurrency models are based. If Microsoft thought there was a problem with their threading model I guess they would call him up since he now works for them.

Re:Yamhill?

[morbid]

An Oxbridge education means nothing but a school tie or a nod towards egalitarianism by taking on a fixed quota of students from state-run comprehensives.

Good, I've got the politics off my chest now.

Re:Yamhill?

[morbid]

...and another thing, doesn't Microsoft donate substantial amounts of money to the Oxbridge universities, or was it just Cambridge?

Re:Yamhill?

[morbid]

Just one more thing:

I don't suppose you've ever looked at the ReactOS kernel? Would I get an education into NT's threading model by looking at that, or is it not the same?

X86-64

[OS24Ever]

x86-64 may be more of a desktop migration point, but there are still plenty of IA64 type applications waiting in the wings from Microsoft, IBM and others.

There is always Linux64.

Not to mention the fact that many a beowolf supercomputer would like to be designed on a Itanium 2. There is one at NCSA from IBM with 800 some IA64 chips. They're just waiting for the Itanium2.

Re:X86-64

[f00zbll]

Here is an honest question, because I don't know the answer and want to know. Are there any 64bit applications for windows 2000 advanced server limited edition on the market? A quick search on google didn't turn up much. Only a bunch of old press releases about 3rd parties working on applications for 64bit, but no actually applications that I can find.

If anyone has first hand experience with 64bit applications on windows, please share your experience. I'm not trolling, just honestly curious about real world deployments of 64bit apps on windows.

Re:X86-64

[Anonymous Coward]

there aren't any. that's why intel don't ship itanic an any quantity - literally in the hundreds from what i hear. if m$ had their 64-bit o/s working however, i think we'd see a different story!

Re:X86-64

[roca]

The reason Itanic didn't ship in quantity is because customers don't want to buy a chip that's slow, hot, costly, and incompatible. Imagine that.

Yikes!

[delta407]

With a faster 3MB on-die L3 cache, 6 instructions/cycle and 6.4GB/s of bandwidth

Not to mention 130 watts of power consumption. And you thought Athlons were hard to cool!

Re:Yikes!

[Octorian]

According to Sun's website, the UltraSparc-III 900MHz Cu chip only has a dissipation of 65 watts.

Re:Yikes!

[Anonymous Coward]

I suppose that low heat dissipation is why my blade requires a pair of 24 cm case fans to keep things cool.

Re:Yikes!

[sk8king]

At the gym, I can pretty much generate 200 watts on a bike for a long time. When the chips start requiring around the 300 watt range I will be totally dependant on the power company. I actually already am, but that is simply because of the wife's hair dryer.

Re:Yikes!

[conway]

Not to mention 130 watts of power consumption. And you thought Athlons were hard to cool!

Its not pretty. The 4-CPU Itanium 1 systems are composed almost entirely of fans on the front to create a sort-of wind tunnel for the damn CPUs.
When you turn the machine on, there's so much interference on fan spinup that all the monitors around you degauss :)

run benchmarks in cache == FAST

[johnjones]

well I'm sorry but all the benchmarks seem to be cache hitters and so run pretty damn fast

real systems are about BANDWIDTH

memory bandwidth/latency is the reason AMD killed the P4 in benchmarks

lets see INTEL go up aganst a SUN on a large oracle DB then I will take notice

really this is where SUN make their money

regards

john jones

Re:run benchmarks in cache == FAST

[sql*kitten]

lets see INTEL go up aganst a SUN on a large oracle DB then I will take notice

Actually, Intel systems do pretty well, indeed, better than Sun running Oracle with a 3000G test database. And they do a good job on transaction throughput too.

Re:run benchmarks in cache == FAST

[afidel]

Well the USIII has lots of cache, so fair comparison. And the Itanium 2 had more than 3X the memory bandwidth vs the USIII. The AMD also suffers from a bandwidth perspective V P4, as up until DDR 333 even PC-800 RDRAM was faster and now they have the PC-1600 RDRAM ready to go. Also the TPC-C is not a synthetic benchmark, it is fairly reliable test of complex DB systems. I am far from an Intel fanboy (I personally hate em for their business practices and the fact that they rely more on marketing then engineering) but I think that in many ways the IA64 is a much better solution than x86-64, bolting on yet more extensions to a crappy core ISA does not make a good ISA.

Re:run benchmarks in cache == FAST

[Jagen]

RDRAM is now at PC1066 not PC1600 ie: they moved the FSB from 400 to 533. DDR is on the way out as well soon, its not vapour, there are chipsets available, will be interesting to see how that shapes up, other problem with RDRAM is the latency and cost or production, the traces have to be very exact due to the high speed the bus runs at.

Re:run benchmarks in cache == FAST

[Anonymous Coward]

You should note the Intel machines only do well in clusters. For 1000 and 3000 GB the only non-cluster machines are UltraSparcs and HP-PA based. Comparing non-clustered RISC machines to clustered Intel's doesn't make sense.

Re:run benchmarks in cache == FAST

[afidel]

You are of course correct on the 1600 vs 1066, I hate numeric dislexia some times.

Re:run benchmarks in cache == FAST

[dmelomed]

it makes sense in the cost sense.

Re:run benchmarks in cache == FAST

[VAXman]

memory bandwidth/latency is the reason AMD killed the P4 in benchmarks

There's definitely some confusion here. P4 (with Rambus) has much better bandwith than Athlon. On any memory bandwidth benchmark (e.g. Sisfot Sandra) the slowest P4 is faster than the fastest Athlon. The 533 MHz bus version of P4 does 4.26 GB/s, while the Athlon only does up to 2.4 GB/s or 2.7 GB/s. Latency-wise they're pretty much equivalent. They're the same using DDR, and 400 MHz Rambus is slower than DDR, and 533 MHz Rambus is faster than Rambus. Of course, for most applications, latency is more important than bandwidth which is why they're closer in overall performance than in memory bandwidth performance.

Re:run benchmarks in cache == FAST

[fliplap]

We should also remember that companies buy Sun for reasons beyond speed. Sun has legendary support, when something goes wrong, or when they make a mistake, it gets fixed == FAST.

Companies buy Sun because if for some awful reason a processor in an E10k dies they don't have to shut the machine down. The machine can be opened, still running, and the processors hot swapped.

The fact is, Sun machines are really really dependable, and thats what companies pay so much more for.

Re:run benchmarks in cache == FAST

[Hammer]

Not only can you hot swap cpu's etc. The latest SunOS^H^H^H^Holaris can be upgraded or patched without reboot!

Re:run benchmarks in cache == FAST

[roca]

Yeah, instead of extending a crappy ISA, with IA64 Intel managed to design a crappy ISA from scratch :-).

Actually Hammer's long mode is a significant cleanup of the x86 instruction set. A number of rarely-used instructions and architectural features (e.g., segments) are removed, new general purpose registers added, various other features regularized (e.g., you can address the low 8 bits of every single register), and you can even ignore the x87 nightmare and use SSE2 as a clean floating point architecture ... it's faster than x87 too. (GCC can target SSE2 instead of x87 now!)

I just like to say "AMD's Hammer will crush Itanium" :-).

Re:run benchmarks in cache == FAST

[(outer-limits)]

I think you are getting confused with the Battle of Hastings here, please try to stay on topic.

about the 64bit chip

[Anonymous Coward]

what is its codename?

Re:about the 64bit chip

[Anonymous Coward]

CRAP

Yet more Intel vapor-hardware

[NetRanger]

I'm glad Intel is really optimistic about their processor. Now they need to either deliver the goods ... or just add more L1 cache and two more direct memory functions and voila , Pentium 5.

It seems that in comparison to finding ways to rev up the clock speed, PC-based innovation in processors has stagnated -- at least as far as those innovations that actually reach the market.

Perhaps I'm just picky.

Re:Yet more Intel vapor-hardware

[Espen Skoglund]

Eh... how, exactly, do you manage to get from Itanium to Pentium5 by just adding some cache? I also don't get your second point; the idea of how changing to a completely new architecture (including the ISA) is comparable to simply ramping up the clock speed. I suppose you must be referring to changes between Itanium and Itanium2, but do you seriously suggest that Intel should introduce new radical changes in between generations of basically the same CPU?

Sure, you could introduce compatible stuff like HyperThreading, but I believe this is scheduled for the generation after Itanium2 (or the one after that). Nobody should really expect the Itanium2 to be much more than an incremental improvement over the Itanium. The Itanium has afterall always been regaded as sort of a developer's beta version of the architecture.

Re:Yet more Intel vapor-hardware

[Anonymous Coward]

The market will never buy into Itanium. I hate to break this to all you guys, but I would expect a Pentium 5 in higher end shit before I'll ever see an Itanium. Why? Because legacy software is what keeps Intel on top. Once you scrap the x86 architecture you're back on a level playing field. In fact, you're WAY behind PowerPC and Ultrasparcs in native application support and you're going to have to play catchup. Either that or do emulation which sucks.

Re:Yet more Intel vapor-hardware

[oldsk8r]

The market will buy into it, indeed it is now. Just not the desktop market. The IA64 is a server/science platform, and that is the sort of market that is using it. You could build a P4 (or P5 when it's out) box that has the hot swap bits that most IA64 systems have, but it would almost cost as much and you wouldn't get the benefits of the IA64, why do you think Intel make different chips ?

Re:Yet more Intel vapor-hardware

[roca]

> The market will buy into it, indeed it is now.

Intel sold two or three thousand Itaniums, total. Pretty small "market".

I'm ashamed of you guys

[Tri0de]

18 posts and NO ONE has asked us to imagine a Beowulf cluster of these?

seriously, on one hand it sure sounds like marketing hype but on the other hand what I want to know is how well it will benchmark on Doom III!

Re:I'm ashamed of you guys

[OS24Ever]

Read further up the chain a bit, I did, you just have to click on the message to see it.

Re:I'm ashamed of you guys

[Anonymous Coward]

I'll imagine a Beowulf cluster of these when winter hits and my heater is busted!

Re:I'm ashamed of you guys

[morbid]

When your heater is busted? A beowulf of these things would make a good oven for a pizzeria.

What about I/O?

[killmenow]

Who cares about how much faster this CPU will go? It doesn't matter that the CPU will perform 10 gazillion instructions per second or whatever because the I/O bus and Memory architecture are the bottlenecks.

Until they get working implementations of a new I/O bus and a memory architecture that gets RAM bandwidth and latency up to a point where it can keep up with the CPU, this will continue to be nothing more than trivia.

Re:What about I/O?

[Anonymous Coward]

Itanium2 still uses old and slow shared bus.

In 4-way configuration, each CPU gets only 1.6GB/s shared I/O and memory bandwidth.
UltraSparc III has 2.4GB/s memory bandwidth + some I/O bandwidth for each CPU.
Sledgehammer will have 5.4GB/s memory bandwidth for each CPU

Re:What about I/O?

[DarkEdgeX]

Uh, with a 4+ MB cache, the I/O bus and memory architecture aren't nearly as big of a hit on performance as they are on our piddly consumer processors with a mere 512KB of L2 cache. (And up until recently, 256KB was the norm!)

Properly written applications that take advantage of the cache (think video encoders that apply multiple filters on content already in the cache, for one example) are going to scream on this architecture.

Cool, but...

[A_Non_Moose]

does it make the intraweb go faster?

Odd.

[Anonymous Coward]

The site lists their source as "VIA" iirc they are not making chipsets for Intel for anything higher than the pentium iii right now. Why would they have info on a to-be-announced processor?

Re:Odd.

[Anonymous Coward]

And why in each benchmark does the speed of the UltraSparc 3 keep changing?
800,750,900,1050 yet the Itanium 2 stays the same speed.

There's another variation on this story.....

[8127972]

At here.

Re:There's another variation on this story.....

[mrfiddlehead]

Right. But it's funny that they don't bother mentionning how the 1GHz Itanium 2 will perform against the 2.2GHz P4. Funny that.

Compilers

[Ted Maul]

I gather that the Itanium philosophy is to transfer the complexity to the compiler. The question is, how good are the compilers now? At the moment, it looks like a real bastard of a job putting together a decent one for any Itanium series. That VLIW stuff looks like it needs to be spot on every time to get the performance (don't do 3 fp ops in a row).

When I can run my C++ through an Itanium compiler and have it come out good, then I'll believe it. Benchmarks? Right.

Re:Compilers

[Junks Jerzey]

I gather that the Itanium philosophy is to transfer the complexity to the compiler. The question is, how good are the compilers now? At the moment, it looks like a real bastard of a job putting together a decent one for any Itanium series. That VLIW stuff looks like it needs to be spot on every time to get the performance (don't do 3 fp ops in a row).

Solid information content: 0
Repetition of things heard elsewhere: 10

Re:Compilers

[BlueGecko]

Actually, current compiler technology can probably optimize quite well for a VLIW architecture. The only catch is that for it to work properly, you will need to profile your code. But with that data, a good compiler that can figure out where to optimize shouldn't be that hard to write.

What I really don't get, though, is why no one is focusing on using JITs with these. It strikes me that this is the ideal platform for a JIT, where it can recode parts of the program on-the-fly based on where the bottlenecks are and so forth. I mean, wasn't this the whole point of a just-in-time compiler in the first place? IBM's Java runtime can rival C++ in speed if it is allowed to run for a reasonable length of time, allowing the code to become truly optimized, and since Intel is targeting this thing in a server environment where applications will run for a similarly long length of time I fail to see why they aren't going that route. This has the additional benefit that as our understanding of how to optimize for VLIW improves, the programs do not need to be recompiled, but instead can immediately get the benefit. (I am fully aware that Itanium is supposed to do some of this type of optimization itself, but current specs are utter crap, to put it lightly.)

Interestingly, Sun's MAJC architecture does exactly that, expecting that a JVM or similar virtual machine will run on top. I have no clue what happened to that chip, but it struck me that it had much better potential to kick ass than Intel's Itanium despite having similar designs precisely because it was designed for a JIT to be on top.

Re:Compilers

[roca]

It's hard to generate decent code for the IA64 so building a good JIT for it requires a very large investment. Furthermore the JIT compiler would probably be quite slow so it would have to run longer or achieve larger speedups for it to pay off.

Although a JIT would be able to discover and exploit behavior patterns that didn't show up until runtime (and therefore not exploitable by a static compiler), it's not a panacea. Lots of programs are unpredictable even down to the level of individual loop iterations. Such programs really need small branch penalties and hardware support for instruction reordering ... which IA64 doesn't have.

Re:Compilers

[Aram Fingal]

You mean they spent a billion+ dollars developing this thing to transfer complexity to the compiler?

Re:Compilers

[lert]

If you want to learn more about compilers for the IA64 - architecture, take a look at this: http://sourceforge.net/projects/open64. This project was originally started bi sgi two years ago afair, but abandoned due to complete lack of interest from the OS community

hth

-lert

ps. you might give it a try with hp's IA-64 Linux Developer's Kit (== an emulator). Enjoy!

McKinley is a fast SOB

[Anonymous Coward]

I have had the chance to work with a McKinley box for a few months now, and it is with no doubt the fastest chip in the West, especially for some applications like public key crypto algorithms.

Indeed, McKinley running at 1 GHz can do a 1024-bit private key operation in 0.2 milliseconds - something well beyond any other existing processor. For high-volume secure electronic transactions, McKinley rules.

Re:McKinley is a fast SOB

[Anonymous Coward]

Wouldn't things like encryption perform better with dedicated hardware acceleration, like network attached SSL or a SSL enabled ethernet card? Last time I checked, most of the big E-commerce sites use hardware acceleration because it performs better and dramatically improves stability.

Re:McKinley is a fast SOB

[conway]

especially for some applications like public key crypto algorithms.

Oh, here we go again with the crypto.

You fail to mention that crypto is the ONLY application where Itanium (1) was any good at -- it has a lot of shift units. Is McKinley going to be the same story? Great crypto performance, really crappy integer / everything else performance?
McKinley had better be better than that, or its going to get the same lukewarm reception as the Itanium.

And they do seem to be having the same difficulties of pushing the clockspeed to decent levels as they did with Itanium. It was only by the C-rev of the original Itanium (the last pre-production beta chip realease) that they achieved 800Mhz, while the original target, I believe, was over 1Ghz.

Its sad, since we all thought McKinley, being desgined more by HP and less by Intel, would have good performance besides SSL.

Re:McKinley is a fast SOB

[Anonymous Coward]

You might like to learn about both crypto and IA64 architecture. McKinley beats all other processors in symmetric crypto all right, where shifts are important. But where it really shines is on asymmetric crypto.

The computationally expensive operation in an RSA private key encryption/decryption is a modular exponentiation - an operation carried out exclusively in integer arithmetic. It is here where McKinley is light-years ahead of its competitors.

Re:McKinley is a fast SOB

[roca]

> For high-volume secure electronic transactions,
> McKinley rules.

For the price of a McKinley you could buy a Pentium 4 and a pile of someone's crypto ASICs, and blow the McKinley away.

Re:McKinley is a fast SOB

[Anonymous Coward]

No you couldn't. Not only are such boards quite expensive but, as far as public cryptography is concerned, they trail behind McKinley. Another problem is that the actual performance of such boards is always significantly less than their theoretical peak, due to the data transfer between the board and the main system. Finally, the performance of a Pentium 4 is pathetic - you would need one of those beasts running at more than 4 GHz to even think about keeping up with a 1 GHz McKinley.

better parallelism

[Anonymous Coward]

Site is slightly slashdotted, and most of the data is in gifs. Here's a fact or 2:
Intel's claimed specint2000 and specfp2000 are both about 1.75x the 800MHz itanium. And this with only a 25% clock speedup to 1GHz.

They claim specint2000 is 1.3x Sun Ultrasparc3 1050MHz, and specfp is 2x.

Unfortunately, there is no indication of what the frequency headroom/scalability might be. The main point of the pentium4 architecture is to scale to 4+GHz. Can we assume anything similar for the itanium?

Re:better parallelism

[afidel]

I think the point of the Itanium's is to go wider not deeper. VLIW combined with the fact that modern CPU's can do most operations much faster than they can retrieve data means that you want to do a fetch on a large data/instruction packet and when it arrives decode it and execute it across many execution units including possibly taking both sides of branches. I think the ultimate goal is to make SMP pretty much transparant as you don't care if the instruction goes to an execution unit on one cpu or another.

Good For Linux

[AstroMage]

A lot of posts have noted that the article doesn't actually give benchmarks, and for various reasons the Itanium-2 may not perform as well as claimed. But consider if it does...

Most scientific heavy-duty work, such as EDA (chip design, etc.) is done on Sun and HP stations running their brand of Unix. Now, if an Intel processor based station starts to perform better than a comparable Sun station, at a much lower price, PLUS you run Linux instead of SunOS or HP-UX, a solution that costs a fraction of a price, and you get the same or better performance- well, you now have a VERY good reason for companies to start using Linux based workstations.

So cheer Intel and AMD on- because it's good for Linux! :-)

SPECint????

[maitas]

I like the part where they said that Itanium 2 has 2x the SPECint performance of the original Itanium, since they never published it!! The SPECint performance for Itanium was so bad that only published SPECfp data!
It's just the same thing that happened when IBM published the SPECint/fp for POWER 4 processors. They only publish the data using 1 processor on the p690, so they run the hole SPEC benchmark suite un the 128MB SRAM cache memory, avoiding using regular DRAM. The easy way to see this is that they never published any SPECrate number, so to avoid showing that they don't scale as all processors start competing for the cache.
Sun USIII 1050MHz is almost 54% faster that USII 750MHz, as anyone can check going to the SPEC page (Sun Blade 1000 Model 1750 against Sun Blade 2050), with a 40% clock speed-up (this 14% increment is due to the compiler). This is exaclty the same processor at a faster clock, while Itanium 2 has more cache and a different architecture that Itanium, so a 1.5x to 2x speedup is less than spectacular, I will said.
For transaction processing, thay don't give any clue to show where they get the info from. While they expect to get the best OLTP number for 4-way systems, I don't think they will be able to surpass the AlphaServer ES45 MoDel 68/1000, which is by far the best 4-way system ever. What's worst, WLIW is know for been a poor performer for OLTP, and a great performer for floating-pont (that's why the only publisehd SPECfp!!). They never published any OLTP benchmark for Itanium (nor SAP, Peoplesoft, ORACLE, or even the raped PTC-C), so you can have an idea of how poor it is...
As of today, Fijutso PrimePower with 128 SPARC processors is the faster OLTP server ever (both SAP and TPC-C numbers!), with IBM p690 a close second for TPC-C and Sun SF15K a close second for SAP SD2-tier. Intel never showed in this kind of performance numbers, and Itanium certainly won't (unless while they keep running Windows).

Re:SPECint????

[Anonymous Coward]

How could this post be moded up ?

Go to spec.org, you'll find all kinds of SpecInt results for Itanium (aka Merced). Those for Itanium 2 (aka McKinley) will appear at launch.

Sun USIII 1050Mhz Spec benchmark is the result of cheating on one test where the compiler has been taught to recognize the benchmark and apply a conversion.

Who Cares?

[Anonymous Coward]

IA64 is stillborn and Intel knows it. Their benefactor, Micro$oft, cant get their act together for either generation of IA64, otherwise Intel would have been shipping millions of these things before now.

When did the original developer editions of IA64 go out? 1998? Micro$oft just can't get it together. And because of that, Micro$oft have essentially prevented Intel shipping this chip in any quantity rather than let it be released to a hungry linux server market.

Intel are indentured servants of Micro$oft, this thing will go nowhere. It's just a press release for the benefit of Wall St.

Itanium 2?

[sharkey]

The first Itanic sank ALREADY?

Not impressed

[Anonymous Coward]

Anyone with inside information in the computer biz knows that Intel is a far more aggressive marketer than even MS, with an even bigger tendency to bend the truth. In other words, the only things you should pay attention to regarding Intel products are 1) the actual price you can buy them for once they're available, and 2) the results of comprehensive, independent benchmarks run by a trustworthy organization.

Anything else is just white noise and should be treated as such.

What about the children?

[Anonymous Coward]

How are you speak of I/O speeds when we still don't know the impact that the GHz speeds have on our children? Shame on you!

Personally...

[smashr]

....I will wait for the Tom's Hardware benchmarks. CPU specs mean nothing to me. Now, if the itaium can get 2000fps in quake3, now that is an accomplishment

Itanium 2 compared to /curren/t SPARC

[Anonymous Coward]

Not only are the SPARC comparisions bogus,
but they're comparing current and /old/
SPARC chips to the yet-unreleased Intel
chip. There is an UltraSPARC IV coming
out next year and a V following soon after...

NY Times Credulous

[xtp]

Similar article in today's NY Times citing the Intel publication. Their article highlights that no comparison is made to recent Pentium-Xeon performance improvements. The article also mentions AMD's Hammer and suggests directly that McKinley may not measure up to the hype when compared with other processors.

The article also mentions that Jack Dongarra - keeper of the Linpack-based 500 fastest computer systems - now shows an Itanium-based cluster at the top of the heap.

Unfortunately for programs that don't run out of the cache, there are three dimensions to computer system performance: processor, memory, and io.
Intel marketing has successfully skewed the common perception to the detriment of a more balanced system viewpoint.

Bad codename for a next-gen processor

[Ilan Volow]

Previous McKinleys haven't fared very well.

If you can profile and recompile the kernel/module

[oliverthered]

There are a lot of compiler problems with the
Itanium/Itanium-2 , infact i think most of the development money went on producing good compiler algorithms for optimisation.

This kind of processor would be great for running OS software because you could compile you kernel/modules/XFree/KDE/postgress etc... with a majic +profile for profiling info, and recompile it with better performance later on. this would mean that diffrerent prople would have the kernel/db software optimised for exactly there needs not for a general peropus implementation.

You are right.

[OS24Ever]

You are right, for windows 2000.

However, under Red Hat Linux 64 there are tons.

And note the release date of Itanium 2. Right along side Windows XP. There are supposedly 64 bit versions of SQL waiting to release with .Net Server 64.

Though honestly most people that ask me about 64 bit computing are Unix (Solaris, AIX, others) wanting to migrate to a less expensive hardware plaform running Linux to replace some lower end Sparc or Power3 boxes.

Though working for IBM I tend to work hard at the Solaris conversion than the AIX ones ;)

Re:You are right.

[f00zbll]

Thanks for that info. What type of 64bit applications are being ported, if you don't mind answering?

Re:You are right.

[OS24Ever]

From what we have been told at this point SQL is the main focus to compete with those high dollar installs of Oracle. I've not heard about Exchange but it can't be very far behind.

But SQL is the main focus to start the new major cash cow to compete with Oracle.

Keep in mind, I work for a hardware vendor though. They don't tell us everything.

you can profile and recompile the kernel/modules

[oliverthered]

This kind of processor would be great for running OS software because you could compile your kernel/modules/XFree/KDE/postgress etc... with a majic +profile for profiling info, and recompile it with better performance later on. this would mean that diffrerent prople would have the kernel/db etc.. software optimised for exactly there needs not for a general peropus implementation.

Re:Open Source like AMD?

[maunleon]

Come on...
I don't want to have to compile my cpu too.

---

MIPS is dominating...

[Steveftoth]

If you consider that every playstation and playstation 2 in the world uses it. They all use the MIPS III ISA (I know the ps2 does, not sure about the ps1, that might be MIPS II). I don't know if they are going to use it again for the PS3, but I would guess so. Something like 100 million playstations (psx+ps2) all use it. Crazy to think, but true.

Re:MIPS is dominating...

[Merlin42]

The embedded market seems to still be competetive.Consoles would definately fall into this catagory. We do seem to be settling on a few ISA's in that arena ie MIPS, ARM, x86, and powerpc. But my understanding is that there are several implementations of each of these so that things are still interesting.And others in this catagory are not completely crushed, and probably won't be any time soon. What I fear (and maybe this is only FUD) is that intel will suck the high end workstation market into its fold and kill competetive innovation there.

Even though intel controls the desktop market pretty tightly they do not control the high end and low end. IMO this is because inteligent people can take their time to make decisions either when designing an embedded device or spending tens-of-thousand$ or million$.

Pfffft... One Gigaherz!

[Dr. Spork]

My chip runs faster than one gigaherz!

No, seriously, it seems that initial releases of the Hammer will have 2X the clock frequency of the McKinleys. I hope Intel includes an "Opteron rating" into the names of the various models just to help us keep things straight!

DDR, no Rambus!

[Dr. Spork]

From the slides, it looks like they intend to use this exclusively with DDR-200, at least around the launch. I think this is a wise move by Intel, and bad news for Rambus!

OT: Deltic?

[Anonymous Coward]

I was curious what deltic was, and what it had to do with spelling, and the only thing a google search had that connected the two was your sig. ;)

Anyway, any chance you could elaborate on what deltic means?

Benchmark? What Benchmark?

[Anonymous Coward]

Since when are Intel's own PowerPoint slides accepted as "benchmark numbers"? These are far too vague to be statistically meaningfull, especially considering they don't come from an independent source. I have no doubts that the Mcikley (Itanium 2) kicks major booty, but I'm rather disappointed in all the hoopla over what's basically a marketing presentation from Intel. Someone post independent benchmark numbers and let us all know about them...don't waste our time with Intel's or Sun's or AMD's own PowerPoint slides.

I'm getting interesting error messages

[Anonymous Coward]

CPU is unworkable or be changed, please resetting

Logo....

[_ph1ux_]

I just think its funny that we have the official Intel (r) company logo for stories relating to Intel.... but we use the Borg Gates logo for anything relating to M$. Funny.

Why Is This Chip Necessary?

[Simonetta]

Could someone please tell me what exactly this chip is projected to be used for? What common applications deliver/generate more wealth to the chip's (or the PC system that this chip will be part of) purchaser compared to the 800MHz Celeron?
What is the current advantage to buying this chip instead of an older chip that is 5-10% of the cost of these ultra-fast state-of-the-art CPUs?

In October 2000 I bought several thousand dollars of Intel stock. That very night the stock price lost 40% of its value; going from $62 per share to the low $40s. It has never recovered and is currently in the upper $20s per share. What is this chip going to do to restore the value of Intel's stock?

I'm serious. Please give me some of your Slashdot insight as to why anyone would want to buy this thing? Will the sales of this thing ever generate the funds needed to recoup the R&D investment (never mind generate enough excitement to actually boost the depressed stock price)?

Re:Why Is This Chip Necessary?

[Anonymous Coward]

The target market for IA64 is high end/mid level servers and true workstations not desktops or low end servers where IA32 is more than enough.

IA32 = high volume with unstable and low unit price/margin
IA64 = lower volume but with more stable and higher unit price/margin

You can make money with either model, as long as you get the volume and pricing balance right. Do you see?

Re:Why Is This Chip Necessary?

[topham]

I'm beginning to think I should start trading stock for a living. Every time I buy stock it drops in value. I figure maybe I should try to Sell-Short and then buy some. (not enough to cover it, but enough to cause the price to drop...).

Or, maybe I should just invest heavily in Microsoft.

Re:AMD, " ALL IN ONE " primary/secondary cache ?

[Anonymous Coward]

3D volume holographic optical storage wants
to replace all system memory with "ALL IN ONE"
memory approach for the future.

http://www.colossalstorage.net

unfortunately, it wont be able to keep your
coffee warm or cook your meals.

Itanium has 3x the I/O of UltraSPARC 3

[Anonymous Coward]

The Itanium 2 FSB is 128bits wide and runs at 400 MHz (i.e. 100 HMz quad pumped).

16Bytes * 100 MHz * 4 = 6.4 GBps

The RAM is quad interleaved DDR 200 -- which also yields 6.4 GBps.

There are multiple PCI-X busses.

This information was published over a year ago. What-sa matta? No internet connection in that cave of yours?

can you say ecache data parity error?

[Chaostrophy]

Sure Sun fixes problems fast. This one only got fixed if you had pull and screamed loud.

Sun did change the error message in later versions of Solaris 8.

I know a large company (you would know the name) who's production R3 system was off line for more than a day, mid week, because of this (crashed, and that trigered the problem that kept them down, thanks for the sucky support, EMC!).

Wrong-O Jim... There are 4 published results

[Anonymous Coward]

First of all, you're completely wrong about the Intel not having published SpecINT results for the Itanium 1.

There are four published results:

http://www.specbench.org/osg/cpu2000/results/res 20 01q3/cpu2000-20010802-00778.html

http://www.specbench.org/osg/cpu2000/results/res 20 01q3/cpu2000-20010730-00776.html

http://www.specbench.org/osg/cpu2000/results/res 20 01q3/cpu2000-20010827-00842.html

http://www.specbench.org/osg/cpu2000/results/res 20 01q4/cpu2000-20010911-00899.html

And as long as we're at it, most of the performance improvement in the last UltraSPARC numbers come from one clever compiler optimization that yields a 27x improvement in one (and only one) of the components of the suite -- but of course you knew that, right?

Re:Wrong-O Jim... There are 4 published results

[maitas]

Yes, I did knew that... but hey, that's the point on showing what you want!
By the way, the SPECint for the Itanium are as low as I expected to be (less than half of the USIII or even the fastest Pentium 4!), that's why the will suck in OLTP. OLTP is marely int, and Itanium is only competitive (not the fastest) on fp, that's why the talk about DSS.
By the way, the tunning that Sun uses for the compiler, is only for SPECfp, it does not work on SPECint, so USIII is a real number, but of course you knew that, right?

Re:Wrong-O Jim... There are 4 published results

[Anonymous Coward]

Please, inform yourself. OLTP is about integer performance, cache hierarchy and system architecture.

Re:Wrong-O Jim... There are 4 published results

[Anonymous Coward]

Bloody brilliant mate!

You admit to trolling with bogus claims, you get called on it, you have no factual defense for your trolling, and now the best that you can do is to try attacking me.

I await your next bold revelation. What will it be?

Sky == blue

Water == wet

I'm all aquiver with anticipation.

Excuse me but you're completely wrong

[Anonymous Coward]

The Itanium 2 has a 6.4 GBps bus structure.

It's 128bits wide and it runs at 400 MHz.

Re:Excuse me but you're completely wrong

[Analog Squirrel]

In the 4 processor system in the "old, shared" bus design, each processor only gets a part of the 6.4GB. According to the arithmetic I learned, 6.4GB/4 = 1.6 GB, which is what was stated...

You can't buy 1GHz chips in all Sun boxes

[Anonymous Coward]

For example, take a look at the very top end Sun Workstation (the Blade 2000).

It "supports" the 1Gig chip but you can't buy it. The fastest CPU that you can get is 900 MHz.

So why is it unfair for Intel to do their CAD benchmarks against the 900 MHz US3?

News? Fluff!

[MidKnight]

Ooh! Look at the pretty Intel Marketing fluff piece! All those projected performance numbers that are "Under Embargo Until 12:01 AM EDT, May 29th, 2002". Give me a friggin break. Call me once they actually ship the processor, there's a proven MB/Backplane at a reasonable cost, and there's someone who will support me if/when it breaks.

On second thought, don't even bother to call me then either. I can currently buy a Sun Enterprise 420R right now. What was the point of the story again?

--Mid

Address the real reason I use Sun over Intel...

[johnlcallaway]

A little something called remote maintenance. Opps, forgot. Remote maintenance that is possible on a dial-up modem. Opps again. Please change that to the ability to rebuild, boot, change eeprom settings, power off and on, remotely using a 9600baud modem if I had to.

Faster is not always better. As a system admin with both NT and Unix systems, my goal is availability and managability first, savings second. Let's face it, I could rebuild a Sun Solaris box remotely with a Palm, a VT-100 emulator, and a cell modem from just about anywhere in the country if I had to.

Why is that important? I can only speak for my company, but being able to do the Sun maintenance from the comfort of our homes/desks is very important to me and my staff. We have equipment in remote locations (over 100 miles away from the office), and not having to drive over there to rebuild a server or install patches saves $100 in expenses, plus takes only 20 minutes instead of all day.

This is not a M$ bashing bit either. If we were using Linux on Intel we would have the same issues. What I need Intel to do is very simple, restore a serial console to the platform. Let me have access to the BIOS from the command line and during startup. Let me power on/power off equipment from the console port too.

Yes...I can run a cardpunch machine too!!!

Re:Address the real reason I use Sun over Intel...

[jpmorgan]

You can get server boards that do that.

Next time you buy an intel server, try getting one with an actual server motherboard in it.

64 bit windows, close but not quite

[marcus]

I do have experience running NT on an alpha.

There's not much to comment on. It worked, slowly. There weren't many apps. Tools were very expensive.

I switched it to linux way, way back in the 2.0.xx era. Since then it has acquired my personal uptime record of 350+ days while working 24/7 as home-office smtp/http/squid/ftp/firewall/etc. box. It's as solid as a rock and completely invulnerable to intel-style buffer overflow exploits and other arbitrary(x86) code execution attacks. I get a hit every other day from some skiddie that doesn't know what an alpha is or is not.

Re:Open Source like AMD?

[Anonymous Coward]

boo

Exactly, the single most important mark

[marcus]

Everyone wants the best performance per dollar.
Always.
Everytime.
Everywhere.

Why has a vendor with balls(willing to post FLOPS/$ or whatever transactions/$) never been born? Somebody must have the highest performing system, who are they? Why don't they show themselves?

Performance improvements are conservative...

[LinuxParanoid]

The main point of the pentium4 architecture is to scale to 4+GHz. Can we assume anything similar for the itanium?

I don't follow this too closely any more, but I would presume they'll get to 2+ GHz, maybe 3 GHz, but probably not 4 GHz. A 4x jump is a lot to ask for without some additional redesign, especially if you are talking 4-way SMP running at those rates.

Given that they're claiming a 2x boost in SPECint2000 and SPECfp2000 from Itanium, on the same .18 micron process, the successor chips (Madison/Deerfield) on .13 micron should get them another 2-3x. Those are due sometime in 2003 I think.

--LP

That's not what you said Sunshine

[Anonymous Coward]

What you said was:

" I like the part where they said that Itanium 2 has 2x the SPECint performance of the original Itanium, since they never published it!! The SPECint performance for Itanium was so bad that only published SPECfp data!"

Now you claim to have know from the beginning that the SPECint data was in fact published. Well thanks for knowingly submitting a false claim.

You then go on to write:
"By the way, the tunning that Sun uses for the compiler, is only for SPECfp, it does not work on SPECint, so USIII is a real number, but of course you knew that, right?"

Yes Sunshine I know that 179.art is part of the FP suite.

What I wrote was:
"And as long as we're at it, most of the performance improvement in the last UltraSPARC numbers come from one clever compiler optimization that yields a 27x improvement in one (and only one) of the components of the suite -- but of course you knew that, right?"

Which is absolutely, positively, 100% true.

Unless of course you don't believe that most of the performance gain comes in the FP suite and that most of that gain is due to 179.art running 27x faster on the 1050 than on the 750.

But oh shame on me for dragging ugly old facts into a discussion with a USFanBoy.

Oh FanBoy....

[Anonymous Coward]

...work on those math skills.

The SPECint number for the Itanium 1 and the US3-1050 are 365 and 610 (respectively).

So the Itanium 1 is not (as you claim)"less than half of the USIII".

And if you go on to do the math, you'll see that if the Itanium 2 comes in at 1.75 the rate of the Itanium 1, then the Itanium 2 will be faster than the US3-1050. Of course if the Itanium 2 actually meets the Intel estimates of a clean factor of two then the US3-1050 will be "slower".

OTOH, being faster than the US3 is hardly a great accomplishment. The the Itanium 2+ had better get the lead out.

Re:A fast CPU , oh wahey , the adrenaline buzz ...

[be-fan]

Its like a bunch of car bores sitting in a bar bullshitting for hours about
webber carbs vs bosch jetronic.
>>>>>>>>>>>
Here's your profile:
You use AOL.
You're VCR blinks 12:00
All you're Windows ME software still has the default settings.
You don't do your own oil changes.
You call a handyman whenever something leaks or breaks or burns.
You've never read the instruction manual to anything you own.
In total, you're a complete marshmellow who doesn't like to get his hands dirty.
Some people like tweaking, whether on cars or computers. More power to them!

fud

[Anonymous Coward]

Boy the sun fud people are out in force today to compete with the intel fud people. This is getting as bad as politics.

I can do any all that on my x86 servers

[Anonymous Coward]

That's what the remote management boards are for.

It's a lot cheaper to add a remote management board to my x86 server than to buy a Sun box.

These features have been available for years from the Teir 1 vendors.

Re:I can do any all that on my x86 servers

[johnlcallaway]

And now I have three vendors to call for support when the damn thing doesn't work ...

Of course, we can talk later about the fact I'll have to reboot 10,000 times if I choose M$.....

SPEC2k Numbers

[Anonymous Coward]

It ain't official until it goes up on the SPEC website but these numbers are very close to what's been predicted by people like Paul DeMone.

Itanium 2:
SPECint: 700 (base) SPECfp: 1350 (base)
US3-1050:
SPECint: 537 (base) SPECfp: 701 (base)

http://biz.yahoo.com/bw/020529/280433_1.html

Re:Pay no attention to community moderation

[morbid]

I'd rather not drop below 0, as I can't be bothered to trawl through all those "howto to (something rude) for Linux Weenies" and other such posts. And another thing, how do you get to post at > 0?

PA-RISC transition to Itanium

[Anonymous Coward]

Newer PA-RISC based servers will be upgradable to Itanium-II by simply the CPU module. Show me a P3-based PC that can be upgraded to P4 without changing the motherboard.

At the same time HP has a lot of experience in dynamic recompilation. There is even a JIT PA-RISC-emulator for PA-RISC which on some benchmarks is 10% faster than running the code natively on the chip. Software transition to Itanium should be a breeze.

Both PA-RISC and the Itanium Product Family will be offered in parallel for quite a while. I must admit that I can't see where the "abruptly killed" part came from. The transition from Alpha is a slightly larger hickup, since customers will be going to HP-UX from Tru64 and VMS, but it all seems to be planned for.

Disclaimer: I work for HP. I don't have access to inside information about the Itanium transition. This is only my personal opinion.

PS: HP tends to suck at actually killing off products. There tends to be support and development done on products that were officially dead ages ago. I have even heard salespeople complaining that it is hard to sell new stuff when 10 or 20 year old equipment is still running fine and being supported.

No, you have one vendor to call ...

[Anonymous Coward]

... and that would be IBM.

The build the HW

They provide the OS and web server (Linux and Apache)

They write the server platform (Java and DB2)

But hey, thanks for playing.

Also,

[Anonymous Coward]

I'd like to dedicate this to the mods. Thank you.