Whats the average computing power

Festus
Festus
Joined: 2 Mar 05
Posts: 4
Credit: 76,493
RAC: 0
Topic 192530

Hi all!

Im doing a bit of publicity work for E@H and i'd like to give people who are not einstein jocks an idea of how powerful the concept of distributed computing can be.

Can one of you official fellas give me a rough estimate of how many TFLOPS E@H is delivering at present on average?

Thanks in advance, Festus

transient
transient
Joined: 3 Jun 05
Posts: 62
Credit: 115,835,369
RAC: 0

Whats the average computing power

Second column from the right, the lower number.

http://einstein.phys.uwm.edu/server_status.php

Festus
Festus
Joined: 2 Mar 05
Posts: 4
Credit: 76,493
RAC: 0

Well, I'm a silly Pratt.

Well, I'm a silly Pratt. Should have thought of that myself..

Festus
Festus
Joined: 2 Mar 05
Posts: 4
Credit: 76,493
RAC: 0

That really is plenty

That really is plenty fast.
If E@H were a supercomputer, it would be at 4th place on the top 500 supercomputers list!

Pooh Bear 27
Pooh Bear 27
Joined: 20 Mar 05
Posts: 1,376
Credit: 20,312,671
RAC: 0

RE: That really is plenty

Message 61597 in response to message 61596

Quote:
That really is plenty fast.
If E@H were a supercomputer, it would be at 4th place on the top 500 supercomputers list!


And that's true life computing power. Those "top ..." computers probably never use their full potential. Just like most home computers that people just use to surf, e-mail, etc. Our computers work, as a team, to solve a problem, find a solution, etc.

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4,307
Credit: 249,801,547
RAC: 33,165

RE: If E@H were a

Message 61598 in response to message 61596

Quote:
If E@H were a supercomputer, it would be at 4th place on the top 500 supercomputers list!


Actually E@H would be higher in the rankings, because the number on the server status page is averaged over 24/7 of a week, whereas what's listed in the Top500 is rather peak performance. I've not digged the DB for potential peak performance for a long time, it should be something between 50 and 100 % more.

On the other hand, the fact that we are using volunteers computers and don't have any control about the client machines forces us to make redundant calculations, so the results we get on Einstein@Home could (theoretically) be achieved with a supercomputer with half the power.

Again, on the other hand, Einstein@Home costs only a fraction of a Supercomputer (and is much more fun, too).

BM

BM

DanNeely
DanNeely
Joined: 4 Sep 05
Posts: 1,364
Credit: 3,562,358,667
RAC: 0

Is it really cheaper in

Is it really cheaper in absolute terms, or only in that you're not paying the elctricity bills for 99% of boxes running the client? This gets really tricky to answer since a large fraction of the work is being done on boxes that would be on for other reasons, and while putting a dollar value on a machine that would be turned off otherwise is easy, getting the cost of a machine running at 100% instead of idle is more complex.

Udo
Udo
Joined: 19 May 05
Posts: 203
Credit: 8,945,570
RAC: 0

RE: Is it really cheaper in

Message 61600 in response to message 61599

Quote:
Is it really cheaper in absolute terms, or only in that you're not paying the elctricity bills for 99% of boxes running the client? This gets really tricky to answer since a large fraction of the work is being done on boxes that would be on for other reasons, and while putting a dollar value on a machine that would be turned off otherwise is easy, getting the cost of a machine running at 100% instead of idle is more complex.

...and the cpus will produce more heat (and the system will therefore consume more power) when running on Einstein instead of being idle...

Udo

Festus
Festus
Joined: 2 Mar 05
Posts: 4
Credit: 76,493
RAC: 0

Lovely Discussion,

Message 61601 in response to message 61600

Lovely Discussion, People!

Thanks for all your feedback!

Alinator
Alinator
Joined: 8 May 05
Posts: 927
Credit: 9,352,143
RAC: 0

RE: Actually E@H would be

Message 61602 in response to message 61598

Quote:

Actually E@H would be higher in the rankings, because the number on the server status page is averaged over 24/7 of a week, whereas what's listed in the Top500 is rather peak performance. I've not digged the DB for potential peak performance for a long time, it should be something between 50 and 100 % more.

On the other hand, the fact that we are using volunteers computers and don't have any control about the client machines forces us to make redundant calculations, so the results we get on Einstein@Home could (theoretically) be achieved with a supercomputer with half the power.

Again, on the other hand, Einstein@Home costs only a fraction of a Supercomputer (and is much more fun, too).

BM

LOL, yep the tens of millons of dollars it takes to play in that sandbox can be a real show stopper for most projects. :-D

I guess it's a good thing BOINC's a lot of fun. ;-)

Alinator

Annika
Annika
Joined: 8 Aug 06
Posts: 720
Credit: 494,410
RAC: 0

Not only fun, you can also

Not only fun, you can also learn loads doing BOINC if you have a look at the software and technology that is used and chat with staff members and other participants. Well, an IT student like me certainly can ;-)

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.