IBM & Intel chew up supercomputing chart, EOL for the Pentium 4

[Asher]

The Top500 supercomputing list had a huge turnover in the past six months, exemplifying a move from supercomputers to clusters of computers:
http://news.com.com/2100-1008_3-1019875.html?tag=fd_top

IBM, Intel chew up supercomputer list

By Stephen Shankland
Staff Writer, CNET News.com
June 22, 2003, 11:55 PM PT

IBM and Intel have edged ahead on a quickly changing list of the 500 fastest supercomputers.
Hewlett-Packard leads the herd with 159 systems on the list, which is set to be released Monday, up from 137, six months ago. But IBM has almost caught up by increasing its numbers to 158 from 131. Meanwhile, the number of systems based on Intel processors--mostly Xeons and a handful of Itaniums--increased from 56 to 119.

Big Blue continues to lead when it comes to the largest fraction of the total computing horsepower. Of the total ability of 375 trillion calculations per second, or teraflops, IBM machines account for 34 percent. More than a tenth of the total comes from the Japanese Earth Simulator, a behemoth with a 36-teraflop rating.

Sun Microsystems, despite a new high-performance computing effort, has only nine systems on the list.

The Top500 list is released twice each year by researchers from Germany's University of Mannheim and the University of Tennessee. And although it always changes, the flux in the new list is higher than usual.

Turnover is high, for one thing. The lowest-ranked system on the new list has a speed of 245 gigaflops, or billion calculations per second. Six months ago, that was enough performance for a comfortable 285th-place ranking.

And there are new varieties of systems arriving, including Cray's new X1, which holds 10 spots starting with No. 112.

A Windows system has made a rare appearance. This cluster of Dell Computer servers with 384 Xeon processors built at the Cornell Theory Center ranked No. 50.

And turning up at No. 362 is an HP system with 64 of Intel's forthcoming Itanium 2 6M Madison processors running at 1.5GHz. The system has a speed of 335 gigaflops.

The No. 8 spot went to an HP system with 1,540 Itanium 2 processors at Pacific Northwest National Laboratory.

But Xeon processors are the main reason for Intel's climb.

The company's jump up the list is largely due to increased performance and lower costs, said Tom Gibbs, director of industry marketing at Intel. Intel deliberately worked on keeping prices down to prompt a massive change in the high-performance computing market.

"The hypothesis we had three years ago was that by altering the price-performance (ratio) in these machines, we would see a dramatic increase in demand," Gibbs said. "A year ago we had three systems on the list."

One trend Gibbs noted is that the number of processors in the Intel systems is declining because of increasing speeds. In the last version of the list, the smallest Intel machine had 128 Xeons. The current one has 92.

Also changing are some of the high rankings. Second and third place were occupied by two identical HP systems called ASCI Q at Los Alamos National Laboratory that now have been combined into one with a 14-teraflop speed. The system uses 8,192 Alpha processors.

A Linux Networx system with 2,304 Xeon processors at sister lab Lawrence Livermore National Laboratory was remeasured at 7.6 teraflops, enough to move it past ASCI White, an IBM machine also at that lab.

The list uses the Linpack benchmark test. Although it's a handy test, it's an imperfect measurement of real-world performance in many cases.

But even more interesting, the last Pentium 4 was released today: 3.2GHz. The next consumer chip to come from Intel will be the Prescott chip (which will likely be the Pentium 5), which is set to debut in late October. AMD won't release another chip until September, the Athlon 64.

But AMD's PR ratings are coming under increasing fire now. Their fastest Athlon, the 3200+, performs like a 2800MHz P4, not a 3200MHz like the name implies.

Some reviews:
http://www.hothardware.com/hh_files/CCAM/p4_32ghz.shtml
http://www.extremetech.com/article2/...1132275,00.asp
http://www.tomshardware.com/cpu/20030623/index.html

From the benchmarks, the P4 is almost without exception ahead of the Athlon 3200. Even the 3GHz is frequently quite a bit faster. In some cases, the P4 is way faster.




I knew AMD's PR ratings were a bad idea.

[Asher]

And from ExtremeTech:







And one thing that gets ignored by most sites, but is quite important (to me, at least) is multitasking performance. Of course, AMD doesn't stand a chance against P4s with HyperThreading...

[Cruddy]

Asher: recompute your figures with price V performance figured in.

Personally, I don't lust after a P4. I lust after an Alpha cluster - totally irrational, I know, because I couldn't use the thing as more than a bookend anyway.

[Asher]

Originally posted by Cruddy
Asher: recompute your figures with price V performance figured in.

Actually, if you look at price vs performance, the 3GHz and below are extremely competitive.

2.8GHz P4 (800MHz): $266
Athlon XP 3200+: $451 (equivalent performance)

Or even a 3.0GHz P4 (800MHz): $399

[Skanky Burns]

Interesting. I may just go over to Intel next upgrade.

[Cruddy]

Thanks Asher. Am currently planning a new upgrade and waiting on funds... so by the time it happens I might just be able to afford to go Intel.

Last time I played with an Intel powered computer was 1988!

[Asher]

I'm genuinely concerned for AMD's future.

Their marketshare has been shrinking, their losses widening, and their performance relative to Intel dropping.

The Athlon 64 has to be much faster than the Athlon XP to remain competitive, but from the leaked previews and benchmarks shown that won't be the case. Even AMD's PR ratings peg the Athlon 64's launch rating at 3400+ -- and we can see here just how badly the 3200+ compares to Intel's offering.

Not to mention the fact that Prescott comes out virtually the same time, and is built for speed and likely to be much faster than the current P4s...

And even on a production cost basis: Intel uses 300mm wafers, AMD's still on 200mm. Intel's on the 90nm process, AMD's on the 130nm process. Prescott will be 80mm² in die area, Athlon 64 will be around 130mm².

And AMD has about 3 months where they're quite a bit slower than Intel and not any cheaper.

[Urban Ranger]

What do you know? This old Anand Tech article has the AMD 3000XP+ definitely is faster than P4 2800, edging it out 13-9 in the benchmarks.

[Asher]

Originally posted by Urban Ranger
What do you know? This old Anand Tech article has the AMD 3000XP+ definitely is faster than P4 2800, edging it out 13-9 in the benchmarks.

You do realize that you're looking at the old 2800:
No HyperThreading, slower 533MHz system bus.

Not fooling anyone, UR.

[Q Classic]

i wonder if intel will do to amd what microsoft did to apple?

somehow, i doubt it, since intel has a few more competitors in the arena.

'tis a pity about amd, but it'll be another year at least before i upgrade, and so i'll give amd that long before deciding where i'll fall for that generation of computers.

[Az]

Yes! Yes
Yes
Yes
Yes
Yes
Yes

Intel is the best for ever and ever and ever!!!!!! Wooohooo They are so great I could kiss them in the ass a hundred times. wait, make that a thousand time! They're immense! Wooohooo!


I think it's time to turn off my P3 500Mhz.

[Asher]

It's not an Intel ass-kissing thread, it's a "neener-neener-neener, I was right!" thread.

I've been telling people for years, AMD is a boom and bust company. It's their history. Investors hate them for it. They make chronically stupid business mistakes, have technical f*ckups which cost them huge marketshare (Athlon 64 should've been out a year ago), and try to cost-cut too much...

Intel will be on top, AMD will consistently oscilate around 10-20% marketshare, and enterprise customers will continue to stay away from AMD because of their inconsistency.

[Asher]

Example of stupid AMD business mistakes:
Pretty much ignoring the mobile market. The mobile market is the only PC market still growing sizably, Intel capitalized on this with the Centrino stuff, and AMD's still horsing around with power-sucking modified Athlons that only work in bulky desktop-replacement notebooks.

They don't seem to have much, if any, foresight.

[Az]

It's not an Intel ass-kissing thread, it's a "neener-neener-neener, I was right!" thread.

[Thue]

Originally posted by Asher

A Windows system has made a rare appearance. This cluster of Dell Computer servers with 384 Xeon processors built at the Cornell Theory Center ranked No. 50.

And turning up at No. 362 is an HP system with 64 of Intel's forthcoming Itanium 2 6M Madison processors running at 1.5GHz. The system has a speed of 335 gigaflops.

The No. 8 spot went to an HP system with 1,540 Itanium 2 processors at Pacific Northwest National Laboratory.

I can't resist (from http://www.hp.com/products1/itanium/...ials/pnnl.html):

PNNL has ordered a $24.5 million HP supercomputer consisting of 1,400 next-generation Intel® Itanium® processors. Once fully operational, the supercomputer should be one of the fastest in the world and the world's most powerful Linux-based supercomputer. The new HP supercomputer would have an expected total peak performance of more than 8.3 teraflops-roughly 8,300 times faster than a current personal computer. Calculations that currently take a month to complete could be done in one day on the new system.

"I'm pleased that the U.S. Department of Energy will now have the world's most powerful Linux-based computer for our pivotal work in biological and environmental research. This new computing power will advance scientific study in a host of areas and keep DOE and our national laboratories on the cutting edge of technology. But more than that, it will allow us to better fulfill our mission to the American people to provide the science needed to solve critical energy and national security problems."

[Q Classic]

how is transmeta doing, anyway?

[Asher]

Thue: Go Itanium!

Speaking of Itanium, Madison launches in a week at 1.5GHz. 50% faster than the current Itanium II, they say.

[Thue]

An interesting development in processor design is the move towards multicore chips and multithreaded cores. FX this article about sun and IBM's future chips should make computer enthusiasts droll, especially the part about the 32-thread niagara in 2005: http://www.eetimes.com/story/OEG20030228S0017

Also of interest is the cell processor to be used in the upcoming ps3: http://www.arstechnica.com/****erdes...ystation3.html (edit: replace "****" with "wank")

And yes, intel is doing 2-way multithreading, which they call hyperthreading. Which is really quite cool and makes me consider buying an intel processor. (If they didn't cost so much compared to low-end AMD processors, I have a low budget)

[Asher]

Upcoming Prescott-based Xeon processors are rumoured to do 4-way multithreading.

Remember that Prescott literally doubled the number of ALUs in the P4 core...so that's going to be used in either Prescott or the Xeon derivations of it.

[Thue]

Originally posted by Q Cubed
how is transmeta doing, anyway?

I haven't heard anything about them for some time, though I tend to follow the tech news. They seem to be strugling to stay afloat.
A few news items about a new processor called astro coming up, but nobody seeming very excited about it.

[Q Classic]

i've been following the news off and on, and things have been quiet from their direction lately, which is why i ask.

pity if they're not doing well, their chip was quite a creative endeavour.