How many Wu's are extracted from a 14 Mb block of data ?
26 Jun 2005 15:21:38 UTC
Topic 189420
(moderation:
Every time one downloads data, one ore more wu's are done with this data.
If wu's for this block of data are exhausted, the client is deleting the old data and downloads a new block*.
In order to see how efficient the WU assigment is, it would be usefull to know how many WU's are extracted from an average block of data and what the maximun and average amount of runs is users have been able to do on their machines***.
*)Recently I've aborted this download, as under the cirumstances it would have taken 4 hours** to finish this (was doing two other downloads at the time).
NB: Don't want to invoke a discussion on times here !
**)I know that I, as modem user have been advised not to run E@H, but still like to (broadband is still not availiable here).
But even broadband users are affected by this, though timewise it would be less noticed.
***) Of couse High end machine will get the highest scores here, but it is equally interesting to see the scores for Mid-Range and Low-spec machines !
How many Wu's are extracted from a 14 Mb block of data ?
)
Well, I don't know about an average, but, Mary Jo gave me a log and I got a single sample from it, and for that one Work Unit Data File she got 13 Results over about 7 days ...
You can see it in the Wiki, if you are interested ...
Up to last weeks data, I have 553 results for Einstein@Home ... working some SQL Magic on my collected data (you are welcome) :)
the spread I have is from 1 to 8 results per Work Unit ID ...
The average number is 3.6, and the histogram is:
[pre]
1 57
2 44
3 48
4 340
5 47
6 12
7 4
8 1
[/pre]
The query to obtain this is
[pre]
SELECT mycount, Count( * )
FROM (SELECT `WorkUnitID` , Count( * ) AS mycount
FROM `BOINCResults`
WHERE `BOINCProject` = 'einstein'
GROUP BY `WorkUnitID`
ORDER BY mycount DESC
) AS TheTable
GROUP BY mycount
[/pre]
RE: Up to last weeks data,
)
Paul,
Many Thanks for doing your Query,
3.6 ave and 8 max on 3.2GHz machines !
What is your resource share for E@H ?
Than 13 max on an unknown cpu, linux or solaris machine.
Good to know, but let's wait for the theoretical posible value !
Regards,
EdH
RE: RE: Up to last weeks
)
My share is 100 for all projects, so, most computers do 25% of time for EAH, My mac is 33% as I only have 3 projects there ...
When LHC@Home is live, my computers drop to 20% ... I tend to spread my wealth about a little bit ... I like BOINC for many reasons, but this is one of my favorite... I can still play the stats game (I think we all do) but with projects more in line with my interests.
My start was with SETI@Home and in many ways my heart wants to do more there, but, reality is that this has little redeaming social value ... I like the Physics projects, and there is one for aritficial life in the wings ... So, if you look at my stats, I am hanging in on most projects, but slowly dropping in SETI@Home standing ... now, if I get a Cray or two, maybe I could hold steady ... :)
I think long term I am going to continure to drift in the direction of projects that have more immediacy, Einstein@Home and LHC@Home for the Physics, Orbit@Home and PlanetQuest for exploration, CPDN as I think we need to start looking at that problem ...
Anyway, it was an interesting question, and it made me start a EAH section in the Wiki ... so, the question is preserved and answered there now too ...
RE: My start was with
)
.. and a AirConditioning and Power Plant to keep them running :-)
a very sensible drift ! like to read that. On CPDN I might shock you, My slogan (or better action point) would be:
"We have to stop looking at it and start doing something about it. The lifestyle in the developped world is beyond sustainable and needs adjustment. We can price ourselves lucky that Global Warming is a slow process, so we have time to work on it, but it would not be fair to leave our offspring and the rest of the world with the created symphoms and the actions to solve them."
Paul, of couse this is not intended for you personal but for all of us.
Kind of, prewriting my CPDN 'User of the week' Climate thoughts here !
Wonder if I wrote it or is it just rewriting of what I have been reading over and over ?
The credits we share here are more valuable to me than the credits from E@H or CPDN !
Regards,
Eric de Haan.
Seven years, seven seas, both of us, but for totally oposite goals, meet in virtual reality. Likely will chat with you again, Eighter way, wish you well !
RE: On CPDN I might shock
)
[soapbox]
agree totally.
The two most urgent BOINC projects, in my personal opinion, are CPDN and O@H.
The irony is in comparing them.
With O@H it will be the cleverness of the developped world, almost certainly led by the USA, that saves the planet in the event that we do spot one of these rocks is actually heading for us.
In contrast, with CPDN it will the the sheer bloody minded stupidity of the very same developped world, currently led by the very same USA, that is creating the problem CPDN is currently checking out.
Both the cleverness and the stupidity are technology led. It would be a sick joke indeed if the biosphere is saved by listening to O@H only to be destroyed by a failure to listen to CPDN.
In both cases, it is strongly positive that work is in hand to check out the facts.
[/soapbox]
~~gravywavy
Gravy, It's hard for me to
)
Gravy,
It's hard for me to comment on this, I don't know O@H (can't find link) but from what you say, it reminds me of a SF movie of some years back.
Sorry, don't know everything, for I am only human !
RE: Gravy, It's hard for
)
O@H is at http://orbit.psi.edu/
currently in pre-alpha and not seeking volunteers till maybe November, when volunteers will be wanted for beta testing.
They will be working out accurate orbits of all the chunks of rock larger than 6km across that have been seen in the sky. When they have doen them, they will go on to 'smaller' pieces, like 1km across. Don't know about you, but I don't fancy wearing a rock 1km across!
More worryingly, the larger rocks will dump so much heat into the ocean that the weather would be screwed up for years. Some scientists say that the dinosaurs were killed by o a similar one 65,000,000 yrs ago and I am not qualified to argue either way. A smaller one might have triggered one of the plagues by changing the climate in the areas where the fleas suddenly took off (good news for fleas, bad for people!).
Yes it is very reminiscent of SF, but the rock that buzzed the moon last year, and we already know one will be coming inside the orbits of our geostationary satellites in April 2029 - these both show that the scenario is not totally off the wall. One every 65,000,000 years means a chance of 1 in 1,000,000 of it happening in a my lifetime or in yours... It probably won't happen, but it is worth checking to make sure.
Apart from my soapbox angle, it is a great BOINC project because the data is already being collected for other reasons. The computing resources haven't been there to compute all the orbits. In the long term, once the backlog is worked through, daily updates will take maybe 10-30mins of each participants computer time, and the rest of the time can be spent running E@H, CPDN, or even [heresy]turning the machine off at night[/heresy].
~~gravywavy
RE: Anyway, it was an
)
Paul,
If you are still reading this thread and are still interested in Q&A pearls of wisdom for inclusion in the Wiki, check out Gravywavy's reply to a question entitled "Time to complete result" by a user named Dan. It's in the Cafe and I know that is somewhere you usually don't spend too much time in :). It struck me as a particularly well written piece of advice for new users with slower machines.
I also wanted to tell you what an excellent and worthwhile job I think you are doing. Keep up the good work.
Cheers,
Gary.
RE: I also wanted to tell
)
I agree for a 110%
Thank you ... :) The
)
Thank you ... :)
The reason I still manage to hange in there is from comments like these ...
One of my biggest problems (daily) is mustering the energy to try to continue (depression is sneaky that way). But, having the occasional hint that somewhere, some time, someone actually got value out of my writing makes it worth while.
You are right, that is a decent write up ... I guess I need to start to work on a couple pages I have been putting off for some time ... I wonder if I could talk gravywavy into asking me for Wiki access so he can help *ME* out ... I crave for good writers (as the book said) "like a jubba freak looking for his next spike ..." (but don't ask me which book it was ... my mind is a mire of disjointed thoughts).
Anyway, now all I have to do is to figure out where it should go! :)