Can Einstein@Home pass the 1 Petaflop (1000 Teraflop) barrier?

The computing power of Einstein@Home has exceeded 950 Teraflops for the first time since the project was begun in 2005. Based on the rate that our computing power has been growing, I am hopeful that Einstein@Home will pass the 1 Petaflop barrier before the end of 2012. Einstein@Home volunteers: please keep your computers running over the holiday season, and please sign up any new ones that you might receive as a gift!

Bruce Allen
Director, Einstein@Home

Comments

dunx
dunx
Joined: 13 Aug 10
Posts: 119
Credit: 53470527
RAC: 0

Overclock.net have added E@H

Overclock.net have added E@H as one of their three "Projects of the month"...

Prepare for blast off, after BGB January 2013 is over...

dunx

Richard Haselgrove
Richard Haselgrove
Joined: 10 Dec 05
Posts: 2139
Credit: 2752740655
RAC: 1458126

RE: 965.7 TFLOPS and

Quote:
965.7 TFLOPS and rising.


970.1 TFLOPS

That calculated/reported figure has been rising at a very steady 1 Tflop/hour since I posted it. Still on course for 22:00 UTC tomorrow?

Mike Davis
Mike Davis
Joined: 3 Apr 05
Posts: 12
Credit: 3026924
RAC: 0

976.8 as of now :)

976.8 as of now :)

city & country school
city & country ...
Joined: 17 Jan 06
Posts: 3
Credit: 3870438
RAC: 0

The only problem is that you

The only problem is that you don't seem to have enough work lined up. I am only a small contributor, with a little less than 700k work units done, but I frequently find that you don't have work for my i7 computer.

Ivailo Bonev
Ivailo Bonev
Joined: 21 Feb 05
Posts: 24
Credit: 40391211
RAC: 9526

RE: The only problem is

Quote:
The only problem is that you don't seem to have enough work lined up. I am only a small contributor, with a little less than 700k work units done, but I frequently find that you don't have work for my i7 computer.

See your Einstein@Home preferences, especially the row "Run only the selected applications", they must be set to "yes" (checked). And "If no work for selected applications is available, accept work from other applications?" must be checked. If you don't have GPU, you can "Run CPU versions of applications for which GPU versions are available" (check it, if it's not).

Alex
Alex
Joined: 1 Mar 05
Posts: 451
Credit: 500137909
RAC: 212795

I agree, there must be

I agree, there must be something with the settings.
I have five systems running here, all set to an additional workbuffer of 0.05 days and they never run dry.

Alex
Alex
Joined: 1 Mar 05
Posts: 451
Credit: 500137909
RAC: 212795

RE: Hallo Vegard!RE: I

Quote:
Hallo Vegard!
Quote:
I rather think that it depends on the conversion factors that are applied, if the numbers make sense or not.

As long as I´m with this project, this is more than 7 years now, the conversionfactor is stable at 1,02*10^-5 (TFLOPS/(Cobblestone/d)]. - for E@H !!! - This I prooved by taking the daily data from the server status page, put them into a graphic and made a least square fit. The correlationfactor R^2 for this is 0,99991, so verry good. But intentionaly the factor is 1e-5 --- 100,000 Cobblestone/d for 1 TFLOPS. I think, the difference is somewhat difficult to rule out here.
The definition of the Cobblestone for the done crunching work, was set up for a better comparison of the different BOINC-projects.

By the way, we will get a new peak crunching power this afternoon. The old maximum was 953,8 @ Dec. 10th 20:35.Now it´s 952,5.

Best whishes for a happy, healthy and wealthy NEW YEAR 2013 !!!!

Kind regards and happy crunching
Martin

Hi Martin,

just to be shure that I understood this right:
a RAC of 50.000 means 0.5 TFLOPS / day? Is there a definition for the 'Cobblestone'?

astro-marwil
astro-marwil
Joined: 28 May 05
Posts: 511
Credit: 402550833
RAC: 1068340

Hallo Alex! RE: a RAC of

Hallo Alex!

Quote:
a RAC of 50.000 means 0.5 TFLOPS / day? Is there a definition for the 'Cobblestone'?


The ratio would be 50.000 Cobblestone per day equals 0.5 TFLOPS (Terra Floatingpoint Operations per Seconde) That´s right. How these FLOPS have to be determined, I don´t know. I believe, one can find the definition somewhere in the literature describing, how to setup a BOINC project.

By the way: We may hit the 1 PFLOPS tomorrow, but at midnight it´s back below this line. But on Thursday it will be well above this line, hopefully. Today we had an increase of 17,1 TFLOPS and we need another 25,2 TFLOPS only.

A happy, healthy and wealthy year 2013 !!!!

Kind regards an happy crunching
Martin

dancer42
dancer42
Joined: 29 Nov 12
Posts: 4
Credit: 415035
RAC: 0

yea seti is down again so I

yea seti is down again so I guess i'm back with a paltry 4.2 terraflop

and I will leave it there until well after you get a petaflop.

good luck guy's it was at .974 pettaflop just before I posted.

Jord
Joined: 26 Jan 05
Posts: 2952
Credit: 5779100
RAC: 0

For the record: 985.5

For the record: 985.5

microchip
microchip
Joined: 10 Jun 06
Posts: 50
Credit: 110896229
RAC: 171591

I've made a challenge over at

I've made a challenge over at BOINCstats called "PetaFLOP Crunch" to celebrate the upcoming 1 PFLOP for Einstein@Home. I think we will reach the 1 PFLOP before the challenge's start (which is 5th jan 2013) but it doesn't matter... Anyone is welcome to participate in the challenge and late entrants are allowed as well.

astro-marwil
astro-marwil
Joined: 28 May 05
Posts: 511
Credit: 402550833
RAC: 1068340

Hallo! It will be hard up to

Hallo!
It will be hard up to reach the 1 PFLOPS today.
At 14:45 we hat 989,2 TFLOPS with an increase of 1.3 TFLOPS/h. In princip that´s enough to reach 10001,1 TFLOPS at midnight (UTC) but very often, like yesterday, between 22:45 and 22:50 there occures a drop of about 5 TFLOPS. So we will have very likely the 1 PFLOPS within the early morning tomorrow.

Kind regrds and happy crunching
Martin

Mad_Max
Mad_Max
Joined: 2 Jan 10
Posts: 153
Credit: 2134792969
RAC: 440417

RE: Hi Martin, I believe

Quote:

Hi Martin,

I believe that the spread in crunching time or the credit/hour is not the crucial point here. I assume that the overall number of floating point operations is (almost) the same for each FGRP2 task, regardless of how long it runs on a certain machine. So, if I know this number and count the jobs per day, I can calculate the overall FLOPS of the project. But does it work like that? Or are the total FLOPS calculated from the CREDIT by using a fixed(!?) conversion number credit/FLOPS? If this is the case, the given FLOPS numbers on the server status page are meaningless (or at least very uncertain), since the credit/FLOPS are different for the 3 subprojects (but would explain the recent increase since FGRP2 is running). It would be nice, if an expert could clarify the method of computing the FLOPS (and thereby explain the HUGE differences of FLOPS numbers on different BOINC webpages, e.g. boincstats.com says 435 TFLOPS at the moment...).

Cheers!

FLOPS calculations are from credits (which should be proportional to the number of calculations in the job) from all users/hosts.
So for now project FLOPS estimate a slightly overrated due high credit value of FGRP2 WUs.
Reason for huge (usual x2) differences of FLOPS numbers on different BOINC webpages is usung different formulas to convert credits/day to FLOPS, which is a lot of confusion with it.
More on this, I wrote in this thread on Rosetta@Home forum: http://boinc.bakerlab.org/rosetta/forum_thread.php?id=6156&nowrap=true#74675
(from this post and to end of thread)

P.S.
New E@H record - 990.9 TFLOPS

David S
David S
Joined: 6 Dec 05
Posts: 2473
Credit: 22936222
RAC: 0

RE: RE: RE: ...

Quote:
Quote:
Quote:
... partially due to Seti's latest problem with downloads getting stuck again.

I don't follow Seti so I'm quite oblivious to whether the project is running normally or not. I do understand that Seti supporters tend to keep large caches and tend to use E@H as a backup. Unless Seti has not been supplying work for a few days, I wouldn't think we would see much effect here yet. Sure, caches might be having the shortfall supplied by E@H but until Seti tasks actually run out, I wouldn't think that more E@H tasks are being crunched yet.

I think there's a fair chance we should be able to declare the Petaflop within the next 36 hours.

SETI has been limping for a while, with a recurrence of the scheduler timeout problem it suffered at the very beginning of November. And around 02:50 UTC this morning, one of their collection of creaky and cranky servers left the party: no work has been sent out or reported since then.

They are also due for scheduled maintenance - probably tomorrow, after the holiday. That means we should get the benefit of the big post-holiday office switch on tomorrow, when their volunteers fetch from backup projects. And even if we miss this opportunity, Berkeley has another outage scheduled for electrical supply work between 4-6 January...

965.7 TFLOPS and rising.


It turned out the download problems I was having were on my end (I had a hosts file specifying all Seti downloads to come from one server; I removed the entry and the backlog cleared in minutes). However, I haven't checked there in a couple of days; it sounds like you're saying there is a new problem I haven't heard of yet. Oh joy.

David

David

Miserable old git
Patiently waiting for the asteroid with my name on it.

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4266
Credit: 244924143
RAC: 16679

FTR: 994.5 BM

FTR: 994.5

BM

BM

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6534
Credit: 284710168
RAC: 110462

996.6 TF !! :-) Cheers,

996.6 TF !! :-)

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Richard Haselgrove
Richard Haselgrove
Joined: 10 Dec 05
Posts: 2139
Credit: 2752740655
RAC: 1458126

RE: 996.6 TF !!

Quote:

996.6 TF !! :-)

Cheers, Mike.


Yes, that was the mark at 22:30 - so I lose the bet I made with myself:

Quote:
I think there's a fair chance we should be able to declare the Petaflop within the next 36 hours.


But we're still climbing (996.9 as I type), and SETI are still struggling, so I'm sure y'all will have cracked it by the time I wake up tomorrow morning.

Donald A. Tevault
Donald A. Tevault
Joined: 17 Feb 06
Posts: 439
Credit: 73516529
RAC: 0

RE: FTR: 994.5 BM Now

Quote:

FTR: 994.5

BM

Now down to 993.6.

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4266
Credit: 244924143
RAC: 16679

Yep, but rising again. There

Yep, but rising again. There were some network problems @UWM which caused some interruption in the daemons' (e.g validators) work, roughly from 21:00 to 21:30 UTC. Needed some manual intervention to get some going again, but all should be working now. Just a little delay...

BM

BM

astro-marwil
astro-marwil
Joined: 28 May 05
Posts: 511
Credit: 402550833
RAC: 1068340

Hallo! Maximum today was

Hallo!
Maximum today was 996,9 TFLOPS at 22:45. Then occured this mysterious drop back by 4.0 TFLOPS at 22:50. Its rising further now. Presumably will arrive at 1000 TFOLPS at 3:25 (all UTC).

Kind regards and happy crunching
Martin

dancer42
dancer42
Joined: 29 Nov 12
Posts: 4
Credit: 415035
RAC: 0

just a thought I am running

just a thought I am running about 4 tflops

and have at least 60 wu pending and while it would not change

the overall computation if say any one with a rac of 10,000 was to do only

validation it would make things more immediate for both the slower systems

and the faster ones as well shortening the time the wu are in the field.

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6534
Credit: 284710168
RAC: 110462

Oh yeah - 997.7

Oh yeah - 997.7 :-)

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5842
Credit: 109390226769
RAC: 35890201

I just reported 202 completed

I just reported 202 completed tasks from 7 hosts with GPUs.

After the 06:35:03 UTC server status update (which was just after all those were reported) the score is now 998.4 TFLOPS.

Looks like 200 extra tasks gives about 0.5 TFLOPS (if the whole 200 made it into the system in time for the update) :-).

EDIT: They must have made it in. The following update was 06:55:02 UTC and the score was little changed at 998.5.

EDIT2: Next update at 07:15:02 UTC. 998.9 TFLOPS.

Cheers,
Gary.

Ivailo Bonev
Ivailo Bonev
Joined: 21 Feb 05
Posts: 24
Credit: 40391211
RAC: 9526

999.2 TFLOPS :)

999.2 TFLOPS :)

Neil Newell
Neil Newell
Joined: 20 Nov 12
Posts: 176
Credit: 169699457
RAC: 0

999.6! Could be the next

999.6! Could be the next update's the one.

I submitted this to Slashdot, who've run the news story.

edit: now 999.9 at 08:37 UTC

juan BFP
juan BFP
Joined: 18 Nov 11
Posts: 839
Credit: 421443712
RAC: 0

Congrats get your beers and

Congrats get your beers and ciggars ready!

Floating point speed (from recent average credit of all users) 1000.0 TFLOPS

Page last updated 3 Jan 2013 8:40:02 UTC

lHj2ixL.jpg

 

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6534
Credit: 284710168
RAC: 110462

Way to go E@H contributors,

Way to go E@H contributors, massive epic effort !! :-) :-)

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

agony
agony
Joined: 2 Jul 07
Posts: 7
Credit: 1397467
RAC: 0

congratulations to all of us

congratulations to all of us for hitting the petaflop.

Data
Data
Joined: 19 Jan 11
Posts: 1
Credit: 356649220
RAC: 910

Congrats us :D

Congrats us :D

Russell
Russell
Joined: 1 Oct 11
Posts: 6
Credit: 14411295
RAC: 0

Good stuff everyone, let's

Good stuff everyone, let's find some pulsars!!

MAGIC Quantum Mechanic
MAGIC Quantum M...
Joined: 18 Jan 05
Posts: 1695
Credit: 1042361031
RAC: 1388641
dancer42
dancer42
Joined: 29 Nov 12
Posts: 4
Credit: 415035
RAC: 0

now can you do 100 petaflops.

now can you do 100 petaflops.

microchip
microchip
Joined: 10 Jun 06
Posts: 50
Credit: 110896229
RAC: 171591

yayyyyy we reached it! :D

yayyyyy we reached it! :D

Henk Haneveld
Henk Haneveld
Joined: 5 Feb 07
Posts: 18
Credit: 14114290
RAC: 316

I hate to put a damper on the

I hate to put a damper on the celebrations but I don't think this is a real record. The RAC that is used to determine the 1 Petaflop barrier is currently inflated with the to high credit values for the FGRP2 results.

If the runtime and credit given for S6LV1 is taken as a baseline then the runtime for FGRP2 should yield about 75 to 80 credits not the current 337 leading to lower RAC numbers.

Richard Haselgrove
Richard Haselgrove
Joined: 10 Dec 05
Posts: 2139
Credit: 2752740655
RAC: 1458126

Sure, the whole thing is just

Sure, the whole thing is just numerology, but it's still fun.

And she's still rising - 1001.4 TFLOPS at 10:30 (at least - the status page is a bit slow to update).

Somebody at slashdot was thinking we ought to be counting in base-2 rather than base-10, and gave us a new target of 1024 TFLOPS to aim for.

Or, if our existing Tflop is base-10 (as I suspect it might be), we'd need 1125.9 of them....

Jord
Joined: 26 Jan 05
Posts: 2952
Credit: 5779100
RAC: 0

Congrats all!

Congrats all!

Logforme
Logforme
Joined: 13 Aug 10
Posts: 332
Credit: 1714373961
RAC: 0

RE: I hate to put a damper

Quote:

I hate to put a damper on the celebrations but I don't think this is a real record. The RAC that is used to determine the 1 Petaflop barrier is currently inflated with the to high credit values for the FGRP2 results.

If the runtime and credit given for S6LV1 is taken as a baseline then the runtime for FGRP2 should yield about 75 to 80 credits not the current 337 leading to lower RAC numbers.


Lets just use FGRP2 as a baseline and increase the S6LV1 and BRP4 credits so we can break the 2 petaflop barrier ! :)

Khangollo
Khangollo
Joined: 17 Feb 11
Posts: 42
Credit: 928047659
RAC: 0

Good that I left my laptop

Good that I left my laptop turned on overnight so we made it! :)

MaU38.gif

microchip
microchip
Joined: 10 Jun 06
Posts: 50
Credit: 110896229
RAC: 171591

RE: I hate to put a damper

Quote:

I hate to put a damper on the celebrations but I don't think this is a real record. The RAC that is used to determine the 1 Petaflop barrier is currently inflated with the to high credit values for the FGRP2 results.

If the runtime and credit given for S6LV1 is taken as a baseline then the runtime for FGRP2 should yield about 75 to 80 credits not the current 337 leading to lower RAC numbers.

There's always a party pooper...

Bojan
Bojan
Joined: 16 Nov 12
Posts: 5
Credit: 3724078
RAC: 0

RE: I hate to put a damper

Quote:

I hate to put a damper on the celebrations but I don't think this is a real record. The RAC that is used to determine the 1 Petaflop barrier is currently inflated with the to high credit values for the FGRP2 results.

If the runtime and credit given for S6LV1 is taken as a baseline then the runtime for FGRP2 should yield about 75 to 80 credits not the current 337 leading to lower RAC numbers.

The same think i here to. On SETI i get in 24/7 30.000/33.000 RAC a day but here i get wit easily RAC of more then 60,000/day. But we get a PTERALOP from a RAC, that somene explain it before how, not from a real CPU/GPU Power. So, any chanche that someone explain it how can we manage to reach a real Pteraflom without a "cheating" of that sort computation?

Thanks and cheers

astro-marwil
astro-marwil
Joined: 28 May 05
Posts: 511
Credit: 402550833
RAC: 1068340

Hallo Henk! RE: The

Hallo Henk!

Quote:
The RAC that is used to determine the 1 Petaflop barrier is currently inflated with the to high credit values for the FGRP2 results.


I think, this is only correct in parts. You neglect the tremendous increase of crunching power of the BRP4 by a factor of 3 at minimum, within the last 4 month, as you can see here.
I feel sure, that BM will correct the earnage of FGRP2 soon. It´s a difficult task, as the crunching times do vary by a factor of 10. I had crunching times at minimum of 0.44 h and maximum of 4.5 h, with a mean value of 2.89 +/- 0,68 h ( derived from 320 tasks). That is much wider than in former projects.

Kind regards and happy crunching
Martin

astro-marwil
astro-marwil
Joined: 28 May 05
Posts: 511
Credit: 402550833
RAC: 1068340

Hallo

Hallo dancer42!

Quote:
now can you do 100 petaflops.


Ok! If we can hold an daily increase of 1% in crunching power, we will reach this within 463 days, that is about 1,25 years. Today we will have an increase of about 1.5%. Sooo, what do we wait for ????? Best you start immediately.

Kind regards and happy crunching
Martin

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6534
Credit: 284710168
RAC: 110462

It's a milestone thing.

It's a milestone thing. Something we can celebrate for it's own sake. In and of itself ..... :-) :-)

The Petaflop target is of course an estimate as we have no sensible way of performing the more exacting types of benchmarks available to supercomputing clusters, say. Being a distributed platform ( that is : a world wide network across a huge variety of hardware and links between ) then such metrics may translate poorly.

I think Bruce Allen has compared E@H's power to that of a dumptruck - the huge yellow ones seen in open cut mining - rather than an F1 racecar. More volume than speed.

The fact remains that it is a computing service which has lent itself to relentless brute force crunching across enormous data sets that simply would not have been processed otherwise. This is using special techniques and strategies pioneered here at E@H. And we have found some diamonds within !! :-)

So we have an 'engine' that is under continual refinement, expansion and improvement. This progress is based upon experience gained and feedback rendered by an active group of very helpful volunteer participants interacting with quite talented and dedicated project staff. That dynamic will serve well for future searches. Roll on Advanced LIGO. :-)

Cheers, Mike.

( edit ) I've always wanted to drive one of those. I know a lady who does, over in the mines in Western Australia. She reckons she has the best job in the world.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5842
Credit: 109390226769
RAC: 35890201

RE: ... Bruce Allen has

Quote:
... Bruce Allen has compared E@H's power to that of a dumptruck ...


So if E@H is a dumptruck, are we going to be 100 dumptrucks in 463 days time? :-).

Anyone want to start a book on when we'll get to 2 dumptrucks? :-).

Cheers,
Gary.

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6534
Credit: 284710168
RAC: 110462

RE: RE: ... Bruce Allen

Quote:
Quote:
... Bruce Allen has compared E@H's power to that of a dumptruck ...

So if E@H is a dumptruck, are we going to be 100 dumptrucks in 463 days time? :-).

Anyone want to start a book on when we'll get to 2 dumptrucks? :-).


OK Gary, so be it! :-)

1 BYDT = 1 BigYellowDumpTruck = 1 Petaflop

I'll go for 2 BYDT's no later than Northern Summer Solstice* !! :-)

and then go for a fleet :

Cheers, Mike.

* June 21st 2013 at 05:04 UTC

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

soft spirit
soft spirit
Joined: 27 Oct 10
Posts: 113
Credit: 5880079
RAC: 0

next stop 2Peta Flops?

next stop 2Peta Flops?

Leonardo Daniel Garcia Martin (ARGENTINA)
Leonardo Daniel...
Joined: 4 Dec 08
Posts: 1
Credit: 4803444
RAC: 0

Einstein@home ya supero el

Einstein@home ya supero el teraflop 1012,9Teraflops!!!!!!!

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3522
Credit: 686043288
RAC: 582127

Hi! We will correct the

Hi!

We will correct the granted credit for FGRP tasks today, so we will then see a decline in the FLOPS count again. So we'll most likely have a chance to celebrate 1 PetaFlops again in the not so distant future :-)

Cheers
HB

geonerd
geonerd
Joined: 8 Nov 04
Posts: 10
Credit: 370024
RAC: 0

RE: I hate to put a damper

Quote:

I hate to put a damper on the celebrations but I don't think this is a real record. The RAC that is used to determine the 1 Petaflop barrier is currently inflated with the to high credit values for the FGRP2 results.

If the runtime and credit given for S6LV1 is taken as a baseline then the runtime for FGRP2 should yield about 75 to 80 credits not the current 337 leading to lower RAC numbers.

IMO, the credit should represent more than raw FLOP count. The Fermi and Arecibo data has proven rather useful from a scientific POV, with numerous observational successes. That ought to count for something.

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4266
Credit: 244924143
RAC: 16679

RE: Somebody at slashdot

Quote:
Somebody at slashdot was thinking we ought to be counting in base-2 rather than base-10, and gave us a new target of 1024 TFLOPS to aim for.

Done.

BM

BM