Processing Time Question

Message boards : Number crunching : Processing Time Question

To post messages, you must log in.

AuthorMessage
Profile JAKware

Send message
Joined: 28 Nov 99
Posts: 8
Credit: 4,983,219
RAC: 1,715
United States
Message 28002 - Posted: 19 Sep 2004, 11:48:39 UTC

I am running Seti@Boinc on an older laptop Compaq Intel Mobile 233 64meg (I know small machine). The issue is this; When it starts working on a WU it goes along very fast, then it starts moving slower. Yesterday I downloaded a new WU and it took about 1.5 hours to get to 51%. In the 17 hours since then it has only progressed 20%. By the time I get done with a work unit it has taken abouit 60 hours. I have even tried re-booting to see if this helps, nothing. Keep in mind that this happens on EVERY work unit.

Any ideas why crunching is so fast then gets progessively slower?

Thank you.
>
ID: 28002 · Report as offensive
Astro
Volunteer tester
Avatar

Send message
Joined: 16 Apr 02
Posts: 8026
Credit: 600,015
RAC: 0
Message 28005 - Posted: 19 Sep 2004, 11:52:18 UTC

I don't know why this happens. My AMD athlonXP took a WU to 60% in 3 minutes then slowed down to taking the full, normal 6 hours.
ID: 28005 · Report as offensive
Profile Mike
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 29584
Credit: 49,119,557
RAC: 17,410
Germany
Message 28010 - Posted: 19 Sep 2004, 12:07:37 UTC

Hi

This is a calculation error in seti client.
It will be fixed later, because its an cosmetic issue.

greetz Mike

ID: 28010 · 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 28095 - Posted: 19 Sep 2004, 19:11:15 UTC

My understanding of the rumor control word is that part of what was done with 4.09 was intended to address this issue. Note, I did not ssay it fixed it, but it should be a little better with 4.09 (note: rumor control can get it wrong).

Just think of it as a gas gague in a car. None of them are linear. After you buy a car you have to drive it for awhile before you know if 1/4 tank is 1/4 of a tank or really 1/2 or 1/2204 left. You know the kind, when it hist 1/4 take you had better be pulling into the station for gas...
<p>
For BOINC Documentaion: Click Me!


ID: 28095 · Report as offensive
Profile Keck_Komputers
Volunteer tester
Avatar

Send message
Joined: 4 Jul 99
Posts: 1575
Credit: 2,464,721
RAC: 1,593
United States
Message 28482 - Posted: 20 Sep 2004, 22:13:19 UTC
Last modified: 20 Sep 2004, 22:13:56 UTC

Each loop through the seti app takes longer than the previous one. The frequency sensitivity increases each time. This is not really a bug just the way the program works.

John Keck -- BOINCing since 2002/12/08 --
ID: 28482 · Report as offensive
Profile FalconFly
Avatar

Send message
Joined: 5 Oct 99
Posts: 394
Credit: 18,053,892
RAC: 0
Germany
Message 28483 - Posted: 20 Sep 2004, 22:18:23 UTC - in response to Message 28482.  
Last modified: 20 Sep 2004, 22:19:11 UTC

The old SETI Client also had (depending on Angle Range) a similar behaviour.

e.g. on VLAR Workunits, I saw it zooming through the first ~8% at light speed, afterwards constant progress far below that.

On normal AR Workunits, first ~7.5% above average speed, ~7.5% to ~78% normal speed, and then accelerating again beyond ~78%.
(all observed on Win32 Clients)

Looking at the Graphics during the first ~60% should show exactly what it does, I reckon it's doing all the (very fast) Pulse/Triplet search, after that it switches to the (intensive and relatively slow) FFT calculations including the best Gaussian shaping.
___________________________________________
<p>Scientific Network : 36200 MHz «» 8204 MB «» 815.0 GB </p>
ID: 28483 · Report as offensive

Message boards : Number crunching : Processing Time Question


 
©2016 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.