No 3GHz Soon and No G5 Laptop

[AidenShaw]

Intel's 3.4 is already a 64-bit chip....

I think he misspoke. 'The industry' is already at 3 GHz - with Intel's 3.4 GHz P4 chips.

What he presumably meant to say is that IBM won't reach 3 GHz for another year. Or perhaps, more generically, that 64 bit processors won't reach 3 GHz until next year.

Note that the Prescott/Nocoma chips *are* 64-bit chips, the x86 Extended architecture features are on the silicon.

It hasn't been turned on in the current Prescott Pentium 4s, but by WWDC Intel will ship Xeon-64 chips with the circuitry enabled.

It's a fair bet that those Xeon-64 chips will run at about the same speeds as the current Prescott chips.

The analysts should know this, so it's probably unlikely that he meant to say "64-bit chips are a year off".

 

[nightcap965]

Mea culpa, mea culpa, mea maxima culpa

*applause*

I am sick of people saying, "Steve lied to us", "Steve promised 3 GHz by this year", etc...

He made a prediction, probably based on the best information he had from IBM at the time.

After three days locked in solitude, seeing no one, speaking to no one save himself and his Maker, Steve Jobs emerged today from his office to speak with reporters. He was unshaven, and was wearing black silk sackcloth instead of his customary mock turtleneck shirt.

"It's true," he said, beating his breast. "I lied. I don't know why I ever thought I could get away with it. I'm deeply sorry for the pain and suffering I have caused."

Jobs announced a new Apple special offer program in expiation. "I must do what I can to ease the grief and disillusionment of those faithful customers who waited so patiently for a 3 GHz G5. We will be selling dual-2.5 GHz PowerMac G5s for $79, plus tax. These systems will come with a 23" Apple Cinema display and an iPod Mini in your choice of colors, at no extra charge." Overcome with remorse and shame, Jobs refused to answer further questions. It is believed he has gone on retreat with an order of cloistered, discalced Buddhist ascetics.

Reaction on MacRumors, a popular news and discussion site, was swift. "That sucks!," said one forum member identified only as "Jiggly". "Michael Dell gave me a Dimension 8300 for $49, and threw in a night with his wife as well. When is Apple gonna realize that the customer is always right?"

 

[~Shard~]

Well there are also Audio/music people like myself who would like to plaster tracks of High power plug-in softsynths (jeez, some of those just EAT processor cycles!) convolution reverbs like Space designer (ditto).
This sort of stuff, that is available now, just keeps on making huge demands of the processor, buses, etc. And th enext genration of audio software?
Only goes to show you can NEVER have enough power!

Yes, this is very true – along with video editing, rendering, etc., audio editing, mixing tracking, etc. is quite processor-intensive as well. Not only do the actual tasks take physical time to perform, but due to the size and complexity of the software being used, this impacts performance as well. So, for many of these Pro users, more power is always a welcome thing! Now if you’re just wanting to use Office, Safari, Mail, iLife, well, you’re probably going to be jus fine with a 1.25 GHz G4 iMac (as an example) for quite a long time!

 

[~Shard~]

Actually, that's one of the problems. Working harder doesn't help that much.

It takes something like 8 weeks from the time a wafer enters the system to the time that finished chips come out the other end. That means that any change, no matter how trivial, takes at least 8 weeks PLUS the time required to make the change. This is true even if you're dealing with known process chemistry.

This favors gathering up a bunch of changes and making them all at the same time. Of course, if two of the changes conflict with each other, you're back at the beginning.

It's simply a very time-consuming task. When you combine this with the fact that the 90 nm process itself is new, untested process and will have its own glitches, it's not surprising that this takes a long time.

I completely agree with you, and I appreciate the efforts involved (I was just being silly in my above post ) Going to the 90nm process has introduced many hurdles and complications, and it is no surprise that the 3 GHz G5s will not be out for some time. Although I am still hoping for sooner rather than later – just ‘cause.

 

[tortoise]

VA Tech won the top honors from Computerworld for their cluster (as well as other scientific awards). Thanks for proving beyond any doubt that you don't have any clue what you're talking about.

I'm still waiting for you to refute a single thing I said. I doubt it will ever happen, because you can't. Media marketing copy does not constitute a refutation of some pretty basic and well established engineering fact, and you obviously don't understand the basic technical details of the subject matter.

I've done extensive supercomputing benchmarks and work on both G5s and Opterons, as have many others. It is well documented in the technical forums. What kind of twit waves content-free marketing copy around as "proof" that all the scientists and engineers are wrong? You don't even have a clue what typical supercomputing loads look like or how they interact with system architectures.

Nobody who works on large clusters thinks the PPCs are good for anything but mult-add codes, and that is a tiny segment of the market. Please show me all the supercomputing cluster folks that are asserting that the G5 is good for STREAM-bound codes, or even non-mult-add floating point codes. Any references? Links?

Go back to playing with Photoshop and leave the supercomputing stuff to the guys that actually do it. You are not technically competent to be making such assertions, and that you use content-free marketing copy to support your position is embarrasing.

 

[jragosta]

I'm still waiting for you to refute a single thing I said. I doubt it will ever happen, because you can't. Media marketing copy does not constitute a refutation of some pretty basic and well established engineering fact, and you obviously don't understand the basic technical details of the subject matter.

I've done extensive supercomputing benchmarks and work on both G5s and Opterons, as have many others. It is well documented in the technical forums. What kind of twit waves content-free marketing copy around as "proof" that all the scientists and engineers are wrong? You don't even have a clue what typical supercomputing loads look like or how they interact with system architectures.

Let's see:

On your side - you're an unknown person whose history no one can check who's claiming to be some kind of expert in the field and you insist that everyone else is wrong.

On my side is ComputerWorld and a number of other organizations that have all given awards to VA Tech - not to mention the VA Tech computer science department itself.

I'd say it's pretty clear which position is stronger.

 

[yuphorix]

Let's see:

On your side - you're an unknown person whose history no one can check who's claiming to be some kind of expert in the field and you insist that everyone else is wrong.

On my side is ComputerWorld and a number of other organizations that have all given awards to VA Tech - not to mention the VA Tech computer science department itself.

I'd say it's pretty clear which position is stronger.

But these awards, first off, don't really mean much in the supercomputer world. Most of them tend to point to the high FLOPS the G5's are able to achieve (We all know IBM processors are known for that), and also the amount of space and money that this system has been able to save.

However, I've not been able to find information on the kind of research done with this set of cluster at all, nor is Virginia Tech known for their research in the supercomputer field. My specialty isn't super computers so I don't know which one of you is right, but it seems like the entire Virginia Tech deal tends to more or less be a publicity thing.

 

[Mav451]

While I agree that he technically mispoke. I think everyone that knows alot of processors knows that the Pentium 4's (with the exception of the "Extreme Edition") is at 3GHz in name only, and that clock for clock it is the slowest "top-of-the-line" chip out there (comparing it to the G5 and the AMD FX-5x).

My question is, with Intel dropping the actual speed from it's chip naming, what is AMD going to do since the Athlon XP 2400+ was named that because it was as fast as an Intel 2.4 GHz chip, etc. Are they going to mimic Intel's new naming system?

That's actually a good question...I'm not sure if AMD will follow, b/c their system as of now is only for the Opteron (1xx, 2xx, 8xx representing single, dual, and quad configurations respectively). Considering the 939's Athlon 64 (128-bit memory controller compared to the 754's 64-bit) still use the "plus" name (e.g. 3800+); I don't think there's going to be a change anytime soon.

 

[arrowhead]

new iMac

True true, but read this statement.

Perhaps they solved those issues ... but I wouldn't be too surprised if they didn't.

Neither did he say that they are not planning for new iMac. So, I still expect to see new iMac introduced very very soon Probably with faster G4 like:

20-inch model with G4 1.50Ghz
17-inch model with G4 1.42 or 1.50 Ghz
15-inch model with G4 1.33Ghz

Also, displays could be at higher resolution. Here I am also assuming that Apple will keep it consistent in terms of making iMac all-in-one machine (no headless).

The same thing for new G4 based PowerBook. Nowaday, when compared to the availability of higher resolution displays for PC laptops, I think that PowerBook needs an upgrade in that area, especially for 17-inch and 15-inch models. That way, 17-inch model can practically provide the display area equivalent to 20-inch, likewise, 15-inch for 17-inch equivalent display area. After all, you can only go bigger physically so much.

 

[jragosta]

But these awards, first off, don't really mean much in the supercomputer world. Most of them tend to point to the high FLOPS the G5's are able to achieve (We all know IBM processors are known for that), and also the amount of space and money that this system has been able to save.

However, I've not been able to find information on the kind of research done with this set of cluster at all, nor is Virginia Tech known for their research in the supercomputer field. My specialty isn't super computers so I don't know which one of you is right, but it seems like the entire Virginia Tech deal tends to more or less be a publicity thing.

Sorry, but universities don't spend $5,000,000 on a 'publicity thing'.

 

[AidenShaw]

Dr. Varadarajan seems to back Itanium, not PPC

I'd say it's pretty clear which position is stronger.

So why does Dr. Varadarajan's company have a "world's fastest cluster" announcement on its homepage - a 19.94 TFLOP cluster of 1024 Itanium 2 systems? (http://www.californiadigital.com/)

It will be interesting to see the next Top500 list - especially since the VAtech cluster might not even be on it at all. (http://www.thinksecret.com/news/virginiatech3.html)

Will there be lots of PPC970 clusters (as you seem to expect), or will Itanium and Opteron continue their rapid growth....

 

[Florida Gator]

The same thing for new G4 based PowerBook.

cough* 13 inch widescreen *cough

 

[dr.Zoidberg]

bravo, bravo

Reaction on MacRumors, a popular news and discussion site, was swift. "That sucks!," said one forum member identified only as "Jiggly". "Michael Dell gave me a Dimension 8300 for $49, and threw in a night with his wife as well. When is Apple gonna realize that the customer is always right?"

Bravo, well said, well said indeed....My new DP 1.8 1gig ram Axc should be here tuesday. Is there any Kool-Aid left?

 

[yuphorix]

Sorry, but universities don't spend $5,000,000 on a 'publicity thing'.

What do you think research in fields do? They bring in publicity, and thus potential customers. We all know the kind of research that goes on in the big name laboratories, although the specifics are unknown, but tell me what is Virginia Tech planning to do with their G5 cluster? And, if a department has the money, why not spend $5mil on a supercomputer, especially if they currently don't have a decently sized cluster? Now, what do you do with it? They'll figure it out later.

Politics can waste alot of money on stuff... you'd be surprised. And there's plenty of politics that go on at Universities.

Why is VAtech replacing their computers already? Isn't the point of a supercomputer to build a legacy system that can aid in research? Now VAtech is deciding to throw away their system and go with Xserve... tell me this isn't a publicity action?

 

[afields]

After three days locked in solitude, seeing no one, speaking to no one save himself and his Maker, Steve Jobs emerged today from his office to speak with reporters. He was unshaven, and was wearing black silk sackcloth instead of his customary mock turtleneck shirt.

"It's true," he said, beating his breast. "I lied. I don't know why I ever thought I could get away with it. I'm deeply sorry for the pain and suffering I have caused."

Jobs announced a new Apple special offer program in expiation. "I must do what I can to ease the grief and disillusionment of those faithful customers who waited so patiently for a 3 GHz G5. We will be selling dual-2.5 GHz PowerMac G5s for $79, plus tax. These systems will come with a 23" Apple Cinema display and an iPod Mini in your choice of colors, at no extra charge." Overcome with remorse and shame, Jobs refused to answer further questions. It is believed he has gone on retreat with an order of cloistered, discalced Buddhist ascetics.

Reaction on MacRumors, a popular news and discussion site, was swift. "That sucks!," said one forum member identified only as "Jiggly". "Michael Dell gave me a Dimension 8300 for $49, and threw in a night with his wife as well. When is Apple gonna realize that the customer is always right?"

LOL! funniest thing I read all day.

 

[maverick13]

I'm still waiting for you to refute a single thing I said. I doubt it will ever happen, because you can't. Media marketing copy does not constitute a refutation of some pretty basic and well established engineering fact, and you obviously don't understand the basic technical details of the subject matter.

I've done extensive supercomputing benchmarks and work on both G5s and Opterons, as have many others. It is well documented in the technical forums. What kind of twit waves content-free marketing copy around as "proof" that all the scientists and engineers are wrong? You don't even have a clue what typical supercomputing loads look like or how they interact with system architectures.

Nobody who works on large clusters thinks the PPCs are good for anything but mult-add codes, and that is a tiny segment of the market. Please show me all the supercomputing cluster folks that are asserting that the G5 is good for STREAM-bound codes, or even non-mult-add floating point codes. Any references? Links?

Go back to playing with Photoshop and leave the supercomputing stuff to the guys that actually do it. You are not technically competent to be making such assertions, and that you use content-free marketing copy to support your position is embarrasing.

http://www.apple.com/education/science/profiles/vatech/architecture.html

Lockhart says several features of the new Power Mac G5 computers were especially attractive for Virginia Tech. “AMD’s Opteron can only execute two double-precision floating-point instructions per clock cycle, which kind of limits system performance,” he notes. The G5 processor has two floating-point units. Thus, unlike the Opteron, it can perform fused multiply-adds, giving four operations per clock cycle and achieving a theoretical limit of 8 gigaflops from a 2 GHz processor. All this is in addition to the G5 processor’s Velocity Engine floating point and integer units.

Adds Varadarajan, “When we were evaluating machines and platforms, I realized that the PowerPC 970 processor would be ideal for us. Its fused multiply-add operation gives it a floating-point performance equal to — if not better than — Intel’s Itanium2 solution. Pretty
quickly, I knew that the Power Mac G5 machines would help us reach the goals we had in mind for our supercomputer.”

Also an interesting link, http://forgetcomputers.com/~jdroz/pages/09.html


• The first module of IBM's Blue Gene/L supercomputer placed 73rd with a sustained performance of 1.4 teraflops. When the full Blue Gene/L is completed in 2005, a system with 128 times as many processors as the current prototype, IBM expects performance of 360 teraflops, enough to top the list by a wide margin.
http://news.com.com/New+blood+joins+supercomputer+ranking/2100-1006_3-5107944.html?part=msnbc-cnet
http://www.research.ibm.com/bluegene/
That's PowerPC 440 -based.

 

[AidenShaw]

VATech was pure publicity

Sorry, but universities don't spend $5,000,000 on a 'publicity thing'.

If it wasn't publicity - why the big panic to get it running for the fall Top500 list?

If you have science to do, a random conference deadline isn't important. If, however, you want publicity to attrach research dollars - the conference is all-important.


If it wasn't publicity - why was it shut down (and is still shut down) just a few months after it was started?

One hint here - a Mac-loving friend who has contacts with Mellanox for a Linux project was told that the VATech cluster was unstable due to the lack of ECC memory, and they wanted to get the ECC Xserves in as soon as possible.

Unfortunately, they dismantled the cluster and sold the PM G5s long before Apple could provide the Xserves. Doesn't that sound like a publicity stunt as well?

 

[benpatient]

oh please. Tortise OWNED you guys and that's the best response you have?

he stated very clearly what the PPC is good at, and the VA guy in that apple promo article doesn't say anything different. Notice he doesn't say what they're going to be using the cluster for, or why they would need to spend 5 million dollars on a cluster that they were going to replace in a few months with a superior cluster that wasn't available at first?

Apple created this story for PR, and it worked quite well. Show me all the other universities and laboratories that have "seen the light" and put in orders for a couple thousand XServes....

....

....

hmn....
....

....

right.

 

[Mav451]

http://forgetcomputers.com/~jdroz/pages/09.html

I've seen quite a few references to that link, actually. And its just as guilty (if not much worse) as Intel with the "Mhz" myth. If I actually bought into that article, and that's a big if, then I'd be inclined to believe that a SINGLE 933 G4 has the same "power" as dual Opteron 246's.

AMD dual 246 Opteron 13,667
Apple 800 MHz G4 Power PC
(e.g. iMac/eMac) 11,450
Apple 933 MHz G4 Power PC 13,400

Hell, maybe even my friend's 3 year old iMac is only SLIGHTLY slower than a pair of AMD's Opteron server chips. Wow. Amazing.

Let that sink in a bit...let that piece of utter absurdity humor you. I had to take a breather for a few minutes, whew.

Wut a joke.

Using MTOPs as a performance indicator is far, FAR worse than the Mhz Myth.

 

[Sun Baked]

What do you think research in fields do? They bring in publicity, and thus potential customers. We all know the kind of research that goes on in the big name laboratories, although the specifics are unknown, but tell me what is Virginia Tech planning to do with their G5 cluster? And, if a department has the money, why not spend $5mil on a supercomputer, especially if they currently don't have a decently sized cluster? Now, what do you do with it? They'll figure it out later.

So I guess the people that think that the PowerPC can power a supercomputer are kidding themselves and wasting money their money and/or that it's just a marketing excercise by Apple. Because a PowerPC supercomputer cannot compete with the x86 and Itanium in the supercomputing world.

I guess that means that Lawrence Livermore National Laboratories is fooling themselves with the PowerPC-based supercomputer they're putting together.

 

[yuphorix]

One hint here - a Mac-loving friend who has contacts with Mellanox for a Linux project was told that the VATech cluster was unstable due to the lack of ECC memory, and they wanted to get the ECC Xserves in as soon as possible.

I totally forgot about ECC memory... I wonder what they were thinking running a cluster without ECC memory!

 

[windowsblowsass]

all right this is some of the most ridicuous things i have ever heard you dont buy$5000000 worth of computers for publicity it just doesnthappen and the idea that this was an apple pr stunt is just ridiculous he day you prove that apple convinced an accredited universitie to pay 5 million for no real reason is the day that nader wins the election

 

[yuphorix]

So I guess the people that think that the PowerPC can power a supercomputer are kidding themselves and wasting money their money and/or that it's just a marketing excercise by Apple. Because a PowerPC supercomputer cannot compete with the x86 and Itanium in the supercomputing world.

I guess that means that Lawrence Livermore National Laboratories is fooling themselves with the PowerPC-based supercomputer they're putting together.

Lets not forget all the Xeon supercomputers that Livermore has. I've never said the PPC wasn't a good processor for supercomputer, it just isn't what people need alot of the time. And the article just goes to prove even more that IBM is good at developing processors to do FP calcs. I've never refuted that.

 

[maverick13]

I've done extensive supercomputing benchmarks and work on both G5s and Opterons, as have many others. It is well documented in the technical forums. What kind of twit waves content-free marketing copy around as "proof" that all the scientists and engineers are wrong? You don't even have a clue what typical supercomputing loads look like or how they interact with system architectures.

Nobody who works on large clusters thinks the PPCs are good for anything but mult-add codes, and that is a tiny segment of the market. Please show me all the supercomputing cluster folks that are asserting that the G5 is good for STREAM-bound codes, or even non-mult-add floating point codes. Any references? Links?

.

Do you have links to back up your claims ... or you do not?

 

[maverick13]

oh please. Tortise OWNED you guys and that's the best response you have?

he stated very clearly what the PPC is good at, and the VA guy in that apple promo article doesn't say anything different. Notice he doesn't say what they're going to be using the cluster for, or why they would need to spend 5 million dollars on a cluster that they were going to replace in a few months with a superior cluster that wasn't available at first?

Apple created this story for PR, and it worked quite well. Show me all the other universities and laboratories that have "seen the light" and put in orders for a couple thousand XServes....

....

....

hmn....
....

....

right.

owned who? With what links? Anybode with some basic knowledge can come here and claim what they will with no references.
How convenient...