S@H Classic vs S@H BOINC: what better to run now?

Message boards : Number crunching : S@H Classic vs S@H BOINC: what better to run now?
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Raistmer
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 16 Jun 01
Posts: 6325
Credit: 106,370,077
RAC: 121
Russia
Message 79683 - Posted: 15 Feb 2005, 21:37:32 UTC

Hi all.

Sorry for probably ressurecting already discussed subject but versions are changing and it would be nice to know what version is faster at this moment?

What % of CPU time is used on BOINC client itself (if one uses common BOINC windows client but in minimized state w/o any graphics)? (and how get maximum science calculations from boinc version?)

Does boinc and classic version make the same signal processing or boinc version now has some improvements?

I'm not interesting in some another advantages like "better control of sharing time, participation in another projects" and so on. Only speed of signal processing does matter in this question.

And another big question: what happens with done WUs from BOINC and Classic versions? Do they merged in one database or boinc just duplicates "classic" work now?
What number of times (average) each WU send to users in both projects? (effectiveness depends of this number too IMHO)
ID: 79683 · Report as offensive
Profile Keck_Komputers
Volunteer tester
Avatar

Send message
Joined: 4 Jul 99
Posts: 1575
Credit: 4,152,111
RAC: 1
United States
Message 79685 - Posted: 15 Feb 2005, 21:48:28 UTC

The main advantage of the BOINC version is that the work is only processed 4 times normally. In the original more copies are generated and sent out as requested, with BOINC you will get a no work available message in the same situation.

In fact there was a post several months ago from one of the on-site developers that mentioned that the BOINC version was completeing more work than classic. Even though classic was returning 4 times as many workunits.
BOINC WIKI

BOINCing since 2002/12/8
ID: 79685 · Report as offensive
Profile Raistmer
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 16 Jun 01
Posts: 6325
Credit: 106,370,077
RAC: 121
Russia
Message 79688 - Posted: 15 Feb 2005, 21:58:07 UTC - in response to Message 79685.  

Thanx for answer

> In fact there was a post several months ago from one of the on-site developers
> that mentioned that the BOINC version was completeing more work than classic.
> Even though classic was returning 4 times as many workunits.
Could you give link on that infomation?

And what about "total repetition"? Does this info mean that all already processed by "classic" WUs will send through boinc now?

SETI apps news
We're not gonna fight them. We're gonna transcend them.
ID: 79688 · Report as offensive
Profile Raistmer
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 16 Jun 01
Posts: 6325
Credit: 106,370,077
RAC: 121
Russia
Message 79689 - Posted: 15 Feb 2005, 21:58:09 UTC - in response to Message 79685.  
Last modified: 15 Feb 2005, 21:58:25 UTC

ID: 79689 · Report as offensive
Profile Paul D. Buck
Volunteer tester

Send message
Joined: 19 Jul 00
Posts: 3898
Credit: 1,158,042
RAC: 0
United States
Message 79707 - Posted: 15 Feb 2005, 22:56:40 UTC - in response to Message 79688.  

> And what about "total repetition"? Does this info mean that all already
> processed by "classic" WUs will send through boinc now?

The rumor is no. They process work on one side or another. But that is just rumor.

ID: 79707 · Report as offensive
Professor Desty Nova
Avatar

Send message
Joined: 17 May 99
Posts: 59
Credit: 579,918
RAC: 0
Portugal
Message 79916 - Posted: 16 Feb 2005, 9:00:05 UTC
Last modified: 16 Feb 2005, 9:00:52 UTC

From BOINC_proj mailling list:


The only thing new in 4.09 is that the jpeg file
(as well as the executable) has a digital signature.
This is to make it work with the latest core client,
which requires that all app_version files be signed.

However, there IS a new SETI@home app around the corner,
with some algorithm changes and support for graphics on
Mac and Unix.


-- David A.




SETI@home classic workunits: 1,985 CPU time: 24,567 hours



Professor Desty Nova
Researching Karma the Hard Way
ID: 79916 · Report as offensive
Profile Mchl
Volunteer tester
Avatar

Send message
Joined: 17 Feb 00
Posts: 35
Credit: 231,170
RAC: 0
Poland
Message 79927 - Posted: 16 Feb 2005, 10:32:19 UTC

In This thread: http://setiweb.ssl.berkeley.edu/forum_thread.php?id=4940
Matt Lebofsky gives some info about BOINC and Classic work.

BOINC@Poland: Polish crossproject team.
http://mchl.republika.pl/boincatpoland/
- Join -
ID: 79927 · Report as offensive
p
Avatar

Send message
Joined: 7 Dec 04
Posts: 106
Credit: 15,334
RAC: 0
Tajikistan
Message 79935 - Posted: 16 Feb 2005, 11:29:39 UTC - in response to Message 79927.  

Its irrelavent which you prefer because soon they will end SETI classic...think about it they wouldnt just change over and cause a lot of hassle if they hadnt got a good reasonable gain to get from it...
<BR>AMD XP3200+
<img src="http://boinc.mundayweb.com/seti2/stats.php?userID=2327&amp;trans=off"><img src="http://petrus.homeftp.org/bws/counter_big.php?id=7828479">
<a href="http://homepage.ntlworld.com/paulandrew.odell/">MY SITE!</a>
ID: 79935 · Report as offensive
Profile Jaaku
Volunteer tester
Avatar

Send message
Joined: 29 Oct 02
Posts: 494
Credit: 346,224
RAC: 0
United Kingdom
Message 79942 - Posted: 16 Feb 2005, 12:08:13 UTC - in response to Message 79935.  

> Its irrelavent which you prefer because soon they will end SETI
> classic...think about it they wouldnt just change over and cause a lot of
> hassle if they hadnt got a good reasonable gain to get from it...

And they are slowly transferring servers over to boinc so even if you can do them faster on classic it will take quite a while to upload!

ID: 79942 · Report as offensive
Profile Raistmer
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 16 Jun 01
Posts: 6325
Credit: 106,370,077
RAC: 121
Russia
Message 79948 - Posted: 16 Feb 2005, 12:17:10 UTC - in response to Message 79942.  

Thanx all for answers.
These infos completely answered on my questions.
BOINC is the right choice now :)

Another question: how to forcedly increase S@H boinc worker process priority under WinNT systems in automated manner? It seems the process is restarted with minimal priority for each new WU.
ID: 79948 · Report as offensive
Profile Paul D. Buck
Volunteer tester

Send message
Joined: 19 Jul 00
Posts: 3898
Credit: 1,158,042
RAC: 0
United States
Message 79962 - Posted: 16 Feb 2005, 14:13:27 UTC - in response to Message 79948.  

> Another question: how to forcedly increase S@H boinc worker process priority
> under WinNT systems in automated manner? It seems the process is restarted
> with minimal priority for each new WU.

There is no need to increase the priority on the application. If there is no other work being done on the machine, increasing the priority will not make the processing go any faster.

The intent is to run the applications using the idle time.

ID: 79962 · Report as offensive
Profile Jaaku
Volunteer tester
Avatar

Send message
Joined: 29 Oct 02
Posts: 494
Credit: 346,224
RAC: 0
United Kingdom
Message 79966 - Posted: 16 Feb 2005, 14:32:23 UTC - in response to Message 79948.  

> Thanx all for answers.
No problem :D

> BOINC is the right choice now :)
If only more people would realise!

ID: 79966 · Report as offensive
Profile Raistmer
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 16 Jun 01
Posts: 6325
Credit: 106,370,077
RAC: 121
Russia
Message 79968 - Posted: 16 Feb 2005, 14:39:01 UTC - in response to Message 79962.  


> There is no need to increase the priority on the application. If there is no
> other work being done on the machine, increasing the priority will not make
> the processing go any faster.

Yes I know that. Idea is to run S@H Classic at background of boinc version to load CPU in "out of work" situation for boinc (and, even more important, in case of some boinc failure). This will still help in WU validation and gives more priority to boinc version in case of new data is available to processing, right? It seems that "classic" version still more stable than boinc one so this approach could give more of valuable results, not?

But with current state of business classic version preempts boinc (although they both run at idle priority).
ID: 79968 · Report as offensive
JAF
Avatar

Send message
Joined: 9 Aug 00
Posts: 289
Credit: 168,721
RAC: 0
United States
Message 79969 - Posted: 16 Feb 2005, 14:41:37 UTC - in response to Message 79683.  

> I'm not interesting in some another advantages like "better control of sharing
> time, participation in another projects" and so on. Only speed of signal
> processing does matter in this question.
>
As to the speed question, currently I see very little difference on my computers between Boinc Seti and Seti Classic (I have two AMD 1900's and a Pentium M 1400). The AMD's average around 4 hours and the Pentium M 3 hours per WU.

In the future, one can expect much longer processing times as the WU data is changed (maybe a different format and/or more data) and different analysis methods. This is one of the reasons Classic is on it's way out.
<img src='http://www.boincsynergy.com/images/stats/comb-912.jpg'>
ID: 79969 · Report as offensive
Profile Saenger
Volunteer tester
Avatar

Send message
Joined: 3 Apr 99
Posts: 2452
Credit: 33,281
RAC: 0
Germany
Message 79972 - Posted: 16 Feb 2005, 15:35:59 UTC - in response to Message 79968.  

> Yes I know that. Idea is to run S@H Classic at background of boinc version to
> load CPU in "out of work" situation for boinc (and, even more important, in
> case of some boinc failure).

Why should you do that?
There are 3 real running projects for boinc S@H, CPDN, PP@H, one in the starting block (E@H) and one on the way over there (LHC).
So no really need for alternatives, one of this projects will provide you with WUs to crunch.
Gruesse vom Saenger

For questions about Boinc look in the BOINC-Wiki
ID: 79972 · Report as offensive
karthwyne
Volunteer tester
Avatar

Send message
Joined: 24 May 99
Posts: 218
Credit: 5,750,702
RAC: 0
United States
Message 79973 - Posted: 16 Feb 2005, 15:44:27 UTC - in response to Message 79968.  

> Yes I know that. Idea is to run S@H Classic at background of boinc version to
> load CPU in "out of work" situation for boinc (and, even more important, in
> case of some boinc failure). This will still help in WU validation and gives
> more priority to boinc version in case of new data is available to processing,
> right? It seems that "classic" version still more stable than boinc one so
> this approach could give more of valuable results, not?
>
> But with current state of business classic version preempts boinc (although
> they both run at idle priority).
>
if you wish to only work for seti, yes that is viable. what works for me (as most of my farm cannot handle CPDN or Einstein WUs) is to use SETI Driver to run as IDLE (newer version, have no idea when it came out, but had to re-download everything due to boinc issue many months ago).
with that setup, Classic gets preempted by BOINC.
granted, after each boinc hiccup, i simply finish out the Classic WU i was on and shut classic back down.

S@h Berkeley's Staff Friends Club
ID: 79973 · Report as offensive
Profile Raistmer
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 16 Jun 01
Posts: 6325
Credit: 106,370,077
RAC: 121
Russia
Message 79982 - Posted: 16 Feb 2005, 16:22:25 UTC - in response to Message 79973.  

> if you wish to only work for seti, yes that is viable. what works for me (as
yes, it's the case :)

> most of my farm cannot handle CPDN or Einstein WUs) is to use SETI Driver to
> run as IDLE (newer version, have no idea when it came out, but had to
> re-download everything due to boinc issue many months ago).
> with that setup, Classic gets preempted by BOINC.

So u suggest to install service-version of boinc, correct? Or what "SETI Driver" is?

> granted, after each boinc hiccup, i simply finish out the Classic WU i was on
> and shut classic back down.

Hm, if I understood right there is too much user attention needed in such way.
ID: 79982 · Report as offensive
karthwyne
Volunteer tester
Avatar

Send message
Joined: 24 May 99
Posts: 218
Credit: 5,750,702
RAC: 0
United States
Message 79994 - Posted: 16 Feb 2005, 16:53:36 UTC
Last modified: 16 Feb 2005, 16:53:48 UTC

hmm, sorry if i was wrong, but an presuming you are using Windows

SETI Driver is an add-on for Classic, it allows you to cache work units. it uses the command line version. you can find it here.

you do not need to turn it back off, that is just what i do (set cache to 0 and raise priority). there are great instructions at the site, but essentially you just put SETI Driver into the folder that has the command-line client and start SETI Driver (add to startup if you like) and it does the rest. as you are primary BOINC, you would probably want to only cache 1-2 units per pc. And by choosing the IDLE setting, it runs at a lower priority than BOINC, low setting runs before BOINC.

hope this helps!

S@h Berkeley's Staff Friends Club
ID: 79994 · Report as offensive
Profile Raistmer
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 16 Jun 01
Posts: 6325
Credit: 106,370,077
RAC: 121
Russia
Message 79998 - Posted: 16 Feb 2005, 17:21:49 UTC - in response to Message 79994.  

> hmm, sorry if i was wrong, but an presuming you are using Windows
yes, u are right :)

> SETI Driver is an add-on for Classic, it allows you to cache work units. it
> uses the command line version. you can find it <a> href="http://www.wakeassoc.com/setidriver/">here.[/url]

Thanx a lot! I just used my own bat-script for caching and didn't know what SETI Driver is. Will use it now :)
ID: 79998 · Report as offensive

Message boards : Number crunching : S@H Classic vs S@H BOINC: what better to run now?


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.