Much slower processing with SETI@home ver 4.5

Message boards : Number crunching : Much slower processing with SETI@home ver 4.5
Message board moderation

To post messages, you must log in.

1 · 2 · 3 · 4 . . . 6 · Next

AuthorMessage
Profile Keith Kennedy

Send message
Joined: 28 May 99
Posts: 149
Credit: 244,165
RAC: 0
United States
Message 31808 - Posted: 1 Oct 2004, 15:57:07 UTC

Has anybody else noticed -- my processing times per work unit have almost doubled since the upgrade to SETI@home ver 4.5. Can anybody explain why?
ID: 31808 · Report as offensive
.
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 410
Credit: 16,559
RAC: 0
Message 31813 - Posted: 1 Oct 2004, 16:21:51 UTC

I don't know exactly if the processing time has doubled, but I noticed that the new WU, I downloaded was 5.09 Mb big instead of usually about 600 kb. And it runs more equally (it this the right word??). Instead of finishing the first about 65 % in 3 minutes, it has finished about 30 % in about 30 minutes.
ID: 31813 · Report as offensive
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 15184
Credit: 4,362,181
RAC: 3
Netherlands
Message 31815 - Posted: 1 Oct 2004, 16:27:27 UTC

I doubt seriously that it was the WorkUnit you downloaded, more like the new Seti Cruncher. Why not give us a copy&paste of your GUI or your transfer window?

As for doubling of processing times, I am still crunching on 4.03, maybe that the 4.05 version went back to Beta basics. In that case, welcome to Beta. :)
----------------------
Jordâ„¢

ID: 31815 · Report as offensive
Profile 1202 Program Alarm
Avatar

Send message
Joined: 16 Jun 99
Posts: 239
Credit: 19,164,944
RAC: 38
United Kingdom
Message 31816 - Posted: 1 Oct 2004, 16:30:40 UTC

4.05 seems to have increased WU crunch time by a couple of hours on my Pentium 4 1.9GHz.


SetiUK - The Offical UK Seti Site - Team Lookers
The Space Directory
Visit Seti.org.uk
SETI News Mailing List


[/url]
ID: 31816 · Report as offensive
.
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 410
Credit: 16,559
RAC: 0
Message 31819 - Posted: 1 Oct 2004, 16:58:51 UTC - in response to Message 31815.  

> I doubt seriously that it was the WorkUnit you downloaded, more like the new
> Seti Cruncher. Why not give us a copy&paste of your GUI or your transfer
> window?
>
> Jordâ„¢
>

I actally KNOW it was the WU as I was watching it download

I have at the moment ONE WU size 6.25 Mb



-
ID: 31819 · Report as offensive
Profile 1202 Program Alarm
Avatar

Send message
Joined: 16 Jun 99
Posts: 239
Credit: 19,164,944
RAC: 38
United Kingdom
Message 31820 - Posted: 1 Oct 2004, 17:01:41 UTC
Last modified: 1 Oct 2004, 17:05:11 UTC

The large files are usually pdb (?) files, the one I downloaded when I updated to 4.05 was just over 5mb.


SetiUK - The Offical UK Seti Site - Team Lookers
The Space Directory
Visit Seti.org.uk
SETI News Mailing List


[/url]
ID: 31820 · Report as offensive
.
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 410
Credit: 16,559
RAC: 0
Message 31823 - Posted: 1 Oct 2004, 17:17:53 UTC - in response to Message 31820.  

> The large files are usually pdb (?) files, the one I downloaded when I updated
> to 4.05 was just over 5mb.
>
What is a pdb file?? I just upgraded to 4.05.
ID: 31823 · Report as offensive
Profile mlcudd
Volunteer tester
Avatar

Send message
Joined: 11 Apr 03
Posts: 782
Credit: 63,647
RAC: 0
United States
Message 31824 - Posted: 1 Oct 2004, 17:26:45 UTC

Hi All,
I don't believe your WU is processing slower, it is actually is running as it is supposed to. One of the fixes in the newer version was the movement of the progress bar in a more linear direction. I think we have been used to seeing the first part of the processing appear to zoom by, and then the program slow down. Now you should see equal prgression to time processed.

Regards,

Rocky
ID: 31824 · Report as offensive
Adrien Seldon
Volunteer tester

Send message
Joined: 3 Apr 99
Posts: 12
Credit: 656,515
RAC: 0
United States
Message 31826 - Posted: 1 Oct 2004, 17:36:31 UTC
Last modified: 1 Oct 2004, 17:41:16 UTC

My processing time on a pentium 4 3.2 went from 3.5 hours on the 4.3 units to 5 hours on the first 4.5 unit I downloaded to 6 hours on the second 4.5 unit I just crunched. Granted my processor is hyperthreaded and I am doing two wu's at once but it slowed way down since the upgrade. Plus, the wu keeps going when it says 100.000% done.
ID: 31826 · Report as offensive
Profile Bruno G. Olsen & ESEA @ greenholt
Volunteer tester
Avatar

Send message
Joined: 15 May 99
Posts: 875
Credit: 4,386,984
RAC: 0
Denmark
Message 31827 - Posted: 1 Oct 2004, 17:41:02 UTC - in response to Message 31823.  

> > The large files are usually pdb (?) files, the one I downloaded when I
> updated
> > to 4.05 was just over 5mb.
> >
> What is a pdb file?? I just upgraded to 4.05.

the pdb file is a debug file


S@h Berkeley's Staff Friends Club ©

ID: 31827 · Report as offensive
Profile Benher
Volunteer developer
Volunteer tester

Send message
Joined: 25 Jul 99
Posts: 517
Credit: 465,152
RAC: 0
United States
Message 31837 - Posted: 1 Oct 2004, 18:44:38 UTC

Hey all,

The .pdb file is..correctly..a debug database file.
It is only used if/when there is a seti worker crash.
(not boinc the manager, seti the worker)

It is then used by a special piece of code to backtrace
and determine the function name where the crash happened.
It then also traces backward to find what function called
the crasher..and what called that...and so on.
----
I have only seen the 5 Meg .PDB myself.
No WU has been larger than approx 360K.
----
WU times. Yea mine increased alot also.
I note the 4.05 worker is 80K larger than the 4.03 worker.exe

I checked the CVS (source code updater) for the worker
and only found the status bar accuracy fix and the more elaborate
graphix fix.

So I'm guessing that work units beginning with 01mr04ab have more
interesting results in them.

I checked the graphics display for my long WU (when it was still calculating while over 100% on the bar)
and it was "finding pulses / triplets"...whereas normal WUs at their
end typically seem to alternate between
"computing fast fourier transform" and "chirping".

=Ben
ID: 31837 · Report as offensive
Heffed
Volunteer tester

Send message
Joined: 19 Mar 02
Posts: 1856
Credit: 40,736
RAC: 0
United States
Message 31860 - Posted: 1 Oct 2004, 20:14:17 UTC - in response to Message 31824.  

> I don't believe your WU is processing slower, it is actually is running as
> it is supposed to. One of the fixes in the newer version was the movement of
> the progress bar in a more linear direction. I think we have been used to
> seeing the first part of the processing appear to zoom by, and then the
> program slow down. Now you should see equal prgression to time processed.

No, they are very definitely slower. Not looking at the progress meter, but the completion times.

ID: 31860 · Report as offensive
Profile 1202 Program Alarm
Avatar

Send message
Joined: 16 Jun 99
Posts: 239
Credit: 19,164,944
RAC: 38
United Kingdom
Message 31863 - Posted: 1 Oct 2004, 20:29:16 UTC

Same here, my completion times on the P4 1.9 have gone from around 5 hours 10 minutes to 7 hours 10 minutes.

SetiUK - The Offical UK Seti Site - Team Lookers
The Space Directory
Visit Seti.org.uk
SETI News Mailing List


[/url]
ID: 31863 · Report as offensive
Profile Woyteck - Boinc Busters Poland
Avatar

Send message
Joined: 3 Jun 99
Posts: 49
Credit: 3,203,845
RAC: 0
Poland
Message 31870 - Posted: 1 Oct 2004, 20:52:50 UTC - in response to Message 31863.  

Mine too, from 2 hours, 50 mins, to about 4 hours and 10 mins on Athlon XP thorton @ 2,2GHz

--
Get up, stand up! Don't give up the fight!

Credits will make everybody feel high! ;-)
ID: 31870 · Report as offensive
Bernard Boldt

Send message
Joined: 5 Aug 00
Posts: 5
Credit: 315,019
RAC: 0
United States
Message 31888 - Posted: 1 Oct 2004, 22:11:55 UTC

I tend to agree the WU has doubled in completion time on my amd 3000 from 2.15 to almost 5 hours. on my amd 2500 it has completed 17% in just under 3 hours.
I have shut them both off - I smell a bug somewhere.
ID: 31888 · Report as offensive
Profile The Gas Giant
Volunteer tester
Avatar

Send message
Joined: 22 Nov 01
Posts: 1904
Credit: 2,646,654
RAC: 0
Australia
Message 31893 - Posted: 1 Oct 2004, 22:22:39 UTC

Oh bugger...I wonder what the completion times will be on my 533MHzP3 running W98SE?
Paul
(S@H1 8000+wu's)

BETA
ID: 31893 · Report as offensive
Petit Soleil
Avatar

Send message
Joined: 17 Feb 03
Posts: 1497
Credit: 70,934
RAC: 0
Canada
Message 31897 - Posted: 1 Oct 2004, 22:33:45 UTC

I would like to mention that it don't change anything as for credits.
It does not matter if WU takes longer to process since the computer
time count and not the amount of WU.
ID: 31897 · Report as offensive
Profile mlcudd
Volunteer tester
Avatar

Send message
Joined: 11 Apr 03
Posts: 782
Credit: 63,647
RAC: 0
United States
Message 31952 - Posted: 2 Oct 2004, 2:47:02 UTC
Last modified: 2 Oct 2004, 12:09:34 UTC

Hello All,
First off I WAS WRONG!, Result times are slower. I have added approx 1.8 to 2.28 hours longer times on all boxes. I wonder also if it is coencidence, but the benchmarks on 2 of my machines are lower than before 4.05.. Just curious

Happy Crunching To all, and I will not be so quick to make assumptions in the future.

Have A Great Day And A Better Tomorrow!


Regards,

Rocky


ID: 31952 · Report as offensive
Heffed
Volunteer tester

Send message
Joined: 19 Mar 02
Posts: 1856
Credit: 40,736
RAC: 0
United States
Message 31961 - Posted: 2 Oct 2004, 3:03:25 UTC - in response to Message 31952.  

> but the benchmarks on 2 of my machines are lower than before 4.05.. Just
> curious

Just a fluke. The project application has no bearing on the benchmark. 4.09 has slightly lower benchmark results due to a bug fix, but 4.05 wouldn't affect benchmarking.

ID: 31961 · Report as offensive
1 · 2 · 3 · 4 . . . 6 · Next

Message boards : Number crunching : Much slower processing with SETI@home ver 4.5


 
©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.