Page 2 of 2
Re: Single computer breaks PetaFLOP barrier
Posted: Fri Jun 13, 2008 11:00 am
by theMASS
alancabler wrote:Greetings
Adam,
Adam A. Wanderer wrote:I wish I knew the results of that test, if it took place.
(Linked in my post...)
http://folding.extremeoverclocking.com/ ... s=&t=48936
here's another link:
http://fah-web.stanford.edu/cgi-bin/main.py
That result (4 Million+ pts) was from just a few days... about 4 days I think, forgot the exact time frame...
When the scores were "fresh" and the stats pages gave more detail, you could see it ramp up and back down
like a bell curve, or a square-wave.
No one who knows anything has said anything-
Dang!
By chance I happened to have the daily stats from "that super computer"
Code: Select all
TimeStamp WUs Points
11/14/2007 1102 1,005,062
11/15/2007 1115 1,022,319
11/16/2007 1124 1,034,460
11/20/2007 1144 1,061,333
11/21/2007 1145 1,061,712
11/22/2007 1721 1,886,719
11/23/2007 2035 2,336,078
11/24/2007 2511 3,016,457
11/25/2007 2935 3,621,864
11/26/2007 3201 3,999,053
12/07/2007 3417 4,310,961
12/09/2007 3418 4,312,401
12/13/2007 3422 4,318,161
12/14/2007 3423 4,319,601
12/29/2007 3425 4,321,247
1/25/2008 3427 4,321,658
Re: Single computer breaks PetaFLOP barrier
Posted: Fri Jun 13, 2008 1:19 pm
by alancabler
Thanks theMASS,
That's great. Hard data trumps swiss-cheese memory.
Looks like the "speed bump" was about 3 million points between 11/21 and 11/26- 2007.
Not bad.
They went into that time slot with a million pts. already, so my guess was totally off.
Hope I didn't spill some beans.
Maybe they'll do it again, and talk some of their buddies with supercomputers into doing it, too.
Re: Single computer breaks PetaFLOP barrier
Posted: Sat Jan 17, 2009 7:12 am
by ashdaw84
Wild-@$$ guess #2:
Fermilab's effort was nothing officialy coordinated, they just had to shut down their accelerator for a few days to sweep out the cigarette butts, or whatever, so their mainframe wasn't busy and they decided (down at the pub) to take 'er out for a spin and see what she'd do...
Nah, that can't be it.
Can it?
I wonder, are you an Ex-pat Aussie?
Love the Guess 2 scenario and who knows? maybe an element of truth?
This might sound very simplistic but, wouldn't it be a small amount of money to power a Super Computer for folding than developing and dropping bombs? Wouldn't it make sense to spend money for extending lives than shortening them?
Re: Single computer breaks PetaFLOP barrier
Posted: Sat Jan 17, 2009 1:08 pm
by Flathead74
ashdaw84 wrote:This might sound very simplistic but, wouldn't it be a small amount of money to power a Super Computer for folding than developing and dropping bombs? Wouldn't it make sense to spend money for extending lives than shortening them?
You're preaching to the choir here, my friend.
I doubt that many folks here are of the mind to be flinging bombs about.
Re: Single computer breaks PetaFLOP barrier
Posted: Sat Jan 17, 2009 5:02 pm
by alancabler
ashdaw84 wrote:Wild-@$$ guess #2:
Fermilab's effort was nothing officialy coordinated, they just had to shut down their accelerator for a few days to sweep out the cigarette butts, or whatever, so their mainframe wasn't busy and they decided (down at the pub) to take 'er out for a spin and see what she'd do...
Nah, that can't be it.
Can it?
I wonder, are you an Ex-pat Aussie?
Love the Guess 2 scenario and who knows? maybe an element of truth?
Native Oklahoman here-
Re: Single computer breaks PetaFLOP barrier
Posted: Thu Jan 22, 2009 3:32 pm
by spazzychalk
Adam A. Wanderer wrote:a decade or two, however, perhaps we'll each have a roadrunner on our desk tops.
http://www.nvidia.com/object/personal_s ... uting.html
Adam A. Wanderer wrote:Testing would generate too many protests, harmless or not.
didnt india set one off underground only a few years ago and no one even knew about it till they annoucned it?
Adam A. Wanderer wrote:But, the owners/operators of these rather expensive machines probably have more "critical" projects in mind
have you seen some of the garbage on BOINC? SeTI, one is even looking for a second pi. another trying to play a row of queens in chess. another one trying to find a new prime #. very critical stuff. there are some useful ones up there, but come on.
Re: Single computer breaks PetaFLOP barrier
Posted: Thu Jan 22, 2009 9:04 pm
by codysluder
spazzychalk wrote:have you seen some of the garbage on BOINC? SeTI, one is even looking for a second pi. another trying to play a row of queens in chess. another one trying to find a new prime #. very critical stuff. there are some useful ones up there, but come on.
Now, now. You don't need to be critical of the choices that other people make. (Aren't you the guy who chooses to believe that you might have done something that causes your computer to overheat?)
A new prime number can lead to significant advances in encryption methods.
Re: Single computer breaks PetaFLOP barrier
Posted: Thu Jan 22, 2009 11:36 pm
by alpha754293
The Roadrunner is a sexy beast!
At cost estimates of between $133-200 M ...yea...they're not going to be running F@H any time soon. Besides, because of their hybrid architecture (AMD Opteron I/O processors, PowerXCell for the computations), if people are having gripes about F@H on Solaris/SPARC as it is, I DOUBT that anything's going to be done to port F@H over specifically for the Roadrunner.
But needless to say, the Roadrunner is scheduled to do MORE nuclear simulations (you know...cuz as in all of the prior ASC computers can't do enough of it between LLNL, LANL, and SNL), some human genome stuff, some geophysics stuff (I forget what exactly), and climate modelling.
(Actually, if you read some of the tech briefs and some of the accomplishments that the Roadrunner has already made, they're simluating stuff at the atomic level, and to think that F@H might be, at best, at the molecular level. HA!)
It never ceases to amaze me though and fascinate me that people already develop programs for this sort of stuff, and how computationally demanding it is.
Personally though, I think that the supercomputer that I still favor the most is the TACC Ranger (currently #6 on the Top500 IIRC). It feels more like "home" to me.
And at ONLY $30M, and achieving a peak of roughly half of what the Roadrunner is capable of, it's a freakin' sweet deal!
Re: Single computer breaks PetaFLOP barrier
Posted: Fri Jan 23, 2009 2:35 am
by spazzychalk
codysluder wrote:(Aren't you the guy who chooses to believe that you might have done something that causes your computer to overheat?)
folding is making my computer overheat. i have this great 1.6ghz dual core laptop and i can only run folding at 40% of cpu because even there its HOT. ive blown it out through the fan and vent side with the heat sink and a lot came out but not enough apparently. no way i trust myself enough to open it up. if it was a desktop sure, but no chance im cracking open a laptop...... but what are you talking about? how coudl ---I--- make the computer overheat?
Re: Single computer breaks PetaFLOP barrier
Posted: Fri Jan 23, 2009 7:11 am
by alpha754293
Adam A. Wanderer wrote:Now that we've entered the computer age, the demand for more and more computing power will never be satisfied, there's just too much it can do.
Well...I don't know if it's so much about "entering the computing age".
As a mechanical engineer, we've often had to simplify our problems LONG before computers were invented just because if we were to actually do it right/properly, the computations would essentially never end.
And as Jonathan Schwartz says, HPC is one of the areas in computing where we will consistently outstrip our demand for computational power, and it is also one of the (few) areas where our demand for it will ALWAYS outpace what Moore's Law allows.
Horray for quantum computing.