TAS members' thread February-March 2007: DPAD and ex-NanoHive@Home.

[caferace]

I'm impressed with the variety, as well.

-jim

 

[GLeeM]

Almost a third of my WUs are erroring. Is this normal?

 

[caferace]

Originally posted by: GLeeM
Almost a third of my WUs are erroring. Is this normal?

Are you running the screensaver? What's the WU name?

-jim

 

[petrusbroder]

Originally posted by: BlackMountainCow

total of 11

Won't happen! Even during the monthly polls we hardly get to 11! You know, we're so 7331, and cool, and elite like, that the others are scared to vote on the Olymp of TAS :shocked:

Well, not quite as bad:

2007 Jan: 11 votes
2006 Dec: no poll
2006 Nov: 10 votes
2006 Oct: 12 votes
2006 Sept: 15 votes
2006 Aug. 13 votes
2006 July: 12 votes
2006 June 13 votes
2006 May: 13 votes
2006 April 13 votes
2006 March 12 votes
2006 Feb. 14 votes
2006 Jan. 15 votes

It is only in November 2006 we had less than 11 votes ... and the lack of polls is explained by the lack of time and energy of the undersigned admin ...

And yes - we are a select group - some 20 active members, with a voting frequence of 50 - 75% not too bad!

And the capacity to produce some 35 000 - over 1 000 000 credits in one single month ... depending on project and number of participants ...

Stats for NanoHive will be done Saturday ...
Poll for NanoHive will be closed @ Saturday

Edit: corrected above numbers (see bold)
Added one line (Stats for ...)

 

[caferace]

Thanks Peter.... I'd be curious how many creds TAS has done cumulatively in a 12 month rolling period. Of course, that takes time, so.......

cheers,

-jim

 

[BlackMountainCow]

Jeez, I guess I'm getting old! I would have sworn we're usually below 11 votes. But now for something completely different!

 

[BlackMountainCow]

Oh, and btw, I finished my 70,000 in Proteins@Home, so that means another X23800+ on Nano! :evil:

 

[BlackMountainCow]

@rise: Hehe, looks like I'm pulling away from your in the ranks ...

:evil:

 

[petrusbroder]

Originally posted by: caferace
Thanks Peter.... I'd be curious how many creds TAS has done cumulatively in a 12 month rolling period. Of course, that takes time, so.......

cheers,

-jim

I thought of doing that again ...

Here are the previous stats - from April 2005 - July 2006, which is 12 months (we did not crunch all the months...)

 

[rise]

Originally posted by: BlackMountainCow
@rise: Hehe, looks like I'm pulling away from your in the ranks ...

:evil:

:laugh:

 

[petrusbroder]

OK, folks, 10 votes are in ...

4 votes for continueing Nanohive@home

4 votes for DPAD
1 vote for Sztaki
1 vote for ABC@Home
6 votes for switching projects

Some more votes please ...

 

[rise]

i don't really see a reason to switch. we are supposed to run the project for a month and aside from the occasional, expected wu error, it runs fine.

which is a good thing because the debacle with the stats has caused me to lose interest really. so its nice it runs pretty smooth and the crunchies get crunched and so on but its the first project (seti aside ) that i haven't cared about running.

so i voted but whatever it is it is.

 

[petrusbroder]

Originally posted by: caferace
Thanks Peter.... I'd be curious how many creds TAS has done cumulatively in a 12 month rolling period. Of course, that takes time, so.......

cheers,

-jim

Well Jim, your wish ...

Here are the accumulated stats for TAS from April 2005 to January 2007
and the stats are just great - take a look.

The running totals are done for the last four months - before that it gets tricky ... since we did quite a few non-BOINC-projects.

I am really proud to be a member of TAS. Not only because TAS has crunched quite a few credits (6.5 Million of them ... ) but because the TAS-crunching team is such a grand group of mates, good friends and a super sub-community in the grand community of Team Anandtech!

 

[BlackMountainCow]

:thumbsup: :thumbsup:

 

[caferace]

Originally posted by: petrusbroder
Well Jim, your wish ...

Here are the accumulated stats for TAS from April 2005 to January 2007
and the stats are just great - take a look.

Thanks Peter. Those are truly amazing. I never knew TAS had that much impact... :thumbsup:

Now go drink. :beer::beer:

-jim

 

[petrusbroder]

Please remember, that we are comparing on a project basis. I.e. TAS' production in one project with TA's production in the same project.
TA crunches - at the same time - quite a few other projects, so the impact of TAS on the total production of TAS is somewhere around 7 - 8%.

 

[caferace]

Ah. Thanks for the explanation. Still, that's a freeway of cobblestones.

Are you drinking yet?

-jim

 

[petrusbroder]

Yepp, I was - but not just now ...

 

[petrusbroder]

A last update and final analysisi on the issue of credit adjustment in NanoHive@Home:

Date ________ number of WUs ______ credit/hour _______ credit/hour ___ credit/hour _____ decrease __
_______________ in stats ___________ average ________ std-dev ______ min - max _____ in credit/hour
February 06 _______ 19 _______________ 48 _____________ 15 ________ 31 - 71 _______ starting point
February 07 _______ 18 _______________ 46 _____________ 23 ________ 21 - 92 __________ -5%
February 08 _______ 17 _______________ 38 _____________ 10 ________ 23 - 52 _________ -20%
February 09 _______ 17 _______________ 34 ______________ 7 ________ 21 - 40 _________ -30%
February 10 _______ 18 _______________ 23 ______________ 9 ________ 13 - 34 _________ -52%
February 11 _______ 12 _______________ 19 ______________ 6 ________ 13 - 27 _________ -60%
February 12 _______ 14 _______________ 22 ______________ 4 ________ 16 - 27 _________ -56%
February 13 _______ 14 _______________ 22 ______________ 5 _________ 9 - 27 _________ -56%
February 14 _______ 18 _______________ 17 ______________ 5 _________ 6 - 24 _________ -65%
February 15 _______ 12 _______________ 17 ______________ 5 ________ 11 - 22 _________ -65%

I think this looks good. Compared to the first numbers the credit/hour has decreased by some 2/3 (from 48 credits/hor to 17 credits/hour).

This is still somewhat higher than some other projects:

Credits/hour ___ Standard ___________ Project
______________ deviation ___________ name
___ 17 __________ 5 ___________ Nanohive@Home (36 WUs run on 20 computers)
___ 17 __________ 4 ___________ Einstein@home (12 WUs run on 7 computers)
___ 10 __________ 3 ___________ Seti@Home (12 WUs run on 8 computers)
___ 10 __________ 6 ___________ QMC@Home (10 WUs run on 6 computers)
____ 6 __________ 1 ___________ Rosetta@Home (12 WUs run on 4 computers - all Macs)
____ 6 __________ 1 ___________ MalariaContol (7 WUs run on 3 slow computers)

I am aware of the fact that seti@home and Einstein@home grant credits as per FLOPS really calculated - but this takes different time on different computers and therefore the credit/hour varies between computers whereas the credits granted are very similar (one WU may take 44 000 seconds on one computer and gets 61 credits, the same WU takes 23 000 second on a faster computer and still gets 61 credits)

 

[BlackMountainCow]

Thx for the update, Peter!

I have to say that I'm not far away from stopping Nano on all of my machines. This stupid popup error box problem is killing me. I'm away from home for the weekend and just checked on my PCs. It looks like all of them have stalles and only on the dualcore and HT machines one "core" is still running. I mean, how can this be? I know, beta, blabla. But they should test their applications never the less. I'll lose a whole weekend of production of my home fleet. The laptop I'm tying this on had 7 out of 13 WUs fail in that way in the last 20 hours! That's almost 50% and that's crap!!!

:|

 

[petrusbroder]

I have stopped new work for NanoHive - because:
1. of the crashing application - have had at least one on each of my comps during the passed 36 hours. It did not stop the crunching - the back-up projects kicked in - but it is annoying.
2. I have 26,981.50 pending credits (99 912.92 granted credits) and it is increasing all the time. I want first this parameter to start dropping before I allow more WUs to be crunched ...

 

[petrusbroder]

OK, folks, 12 votes are in ...

4 votes for continueing Nanohive@home

6 votes for DPAD
1 vote for Sztaki
1 vote for ABC@Home
8 votes for switching projects

Some more votes before midnight?

 

[caferace]

I too had some crashes overnight. Interesting that on my "alienware" box, there is an app called "alien autopsy" that actually dismisses the pop up automatically in favor of it's own pop up. That allows BOINC to resume work, waiting for the next failure. About the only thing the alienware box has going for it, I must say...

I'm ready to switch, simply because I have an embarrassing amount of credits. My pending is 80,171.51 with granted at 230,948.04. More than I ever expected.

Since I've never run DPAD, I'll experiment some today. It's Saturday, freakishly beautiful for this time of year outside and I have no responsibilities today. Maybe a refreshing beverage is in order. Yay!

-jim

p.s... "Beautiful" means ~ 73F and light winds from the NW today. My apologies to those of you (pretty much everyone) in colder climes. If you'd like to drop by, I'm buyin.

 

[caferace]

BTW, Peter? Congrats on breaking 100k! :music::beer::music:

-jim

 

[petrusbroder]

Thanks Jim! :beer:

I am switching some of my boxes to DPAD today - and as they run out of work - some more. I do not know how much I'll add ... I think I'll double or tripple my production though ... 45 000 Mpts/day is not too bad a goal.

Edit:
Well, we have some wonderful sunshine too, the snow is wonderful white, the temp is -4 degrees C, a very comfy winter day. I have been out with the kids for some 3 hours today and that was great - we did some skating (16 Km at a leisurely pace) on the river --- almost no wind, and when there was some, it was gentle and from the side ...
The hot choclate drink was just great!