Version 3.20 on Linux

Message boards : Number crunching : Version 3.20 on Linux
Message board moderation

To post messages, you must log in.

AuthorMessage
Darren
Volunteer tester
Avatar

Send message
Joined: 2 Jul 99
Posts: 259
Credit: 280,503
RAC: 0
United States
Message 6736 - Posted: 12 Jul 2004, 1:21:39 UTC

I updated one of my hosts today to the latest self-compile (boinc 3.20, seti 3.10), which seems a little better and a little worse at the same time.

On the better side, it benchmarked with a better whetstone. P4 3.06 1024 RAM with whetstone of 917 as compared to 813 on self-compiled version 3.18 (compiled from the June 24 source code). Also on the better side, while it was finishing the 2 work units already in the cache, boincgui was reporting the "to completion" times correctly, instead of running it all the way up to 50% done in the first minute, then taking a few hours to do the other 50%. These work units, of course, ran using seti 3.08 though.

Now that it's gone on to a work unit downloaded after the upgrade and processing with seti 3.10, none of the gui add-ons show the progress (they read just like predictor units do on linux gui add-ons (such as "to completion 0-21467484.....").

Lastly, the work units themselves seem to be processing a slight bit faster. It could have just been a fluke and I had 2 that happened to really be just a bit shorter, but since the upgrade the 2 work units that have finished were done in 2:58 and 3:00 - normally this system takes about 3:20 per work unit.

Anyone else do the upgrade and seeing anything similar, or am I the first guinea pig here?


ID: 6736 · Report as offensive
Darren
Volunteer tester
Avatar

Send message
Joined: 2 Jul 99
Posts: 259
Credit: 280,503
RAC: 0
United States
Message 6807 - Posted: 12 Jul 2004, 5:18:30 UTC

Well, I guess I am the only guinea pig here.

Just an FYI for anyone thinking about upgrading, the low-water mark feature doesn't seem to be working in this version. It does access the preferences and reports that it is using them, but a request for more work is only made when the last work unit finishes.

And I guess it was just a 2-shot streak of luck with the 2 work units that finished in 3 hours. The next one took - just like normal - 3 hours and 20 minutes.


ID: 6807 · Report as offensive
Alex

Send message
Joined: 26 Sep 01
Posts: 260
Credit: 2,327
RAC: 0
Canada
Message 6822 - Posted: 12 Jul 2004, 6:17:23 UTC - in response to Message 6807.  

I'm going to assume they made a couple of fixes with the latest version.
They might have some open bug reports for less critical features.

Hopefully they fixed a few bugs related to 'Preferences'
(ie.. client looks at General Prefs instead of Work, Detaching from Projects, etc)
ID: 6822 · Report as offensive
WerK

Send message
Joined: 30 Jun 02
Posts: 26
Credit: 221,390
RAC: 0
Czech Republic
Message 6881 - Posted: 12 Jul 2004, 10:28:34 UTC - in response to Message 6736.  

> I updated one of my hosts today to the latest self-compile (boinc 3.20, seti
> 3.10), which seems a little better and a little worse at the same time.
>
> On the better side, it benchmarked with a better whetstone. P4 3.06 1024 RAM
> with whetstone of 917 as compared to 813 on self-compiled version 3.18
> (compiled from the June 24 source code). Also on the better side, while it
> was finishing the 2 work units already in the cache, boincgui was reporting
> the "to completion" times correctly, instead of running it all the way up to
> 50% done in the first minute, then taking a few hours to do the other 50%.
> These work units, of course, ran using seti 3.08 though.
>
> Now that it's gone on to a work unit downloaded after the upgrade and
> processing with seti 3.10, none of the gui add-ons show the progress (they
> read just like predictor units do on linux gui add-ons (such as "to completion
> 0-21467484.....").
>
> Lastly, the work units themselves seem to be processing a slight bit faster.
> It could have just been a fluke and I had 2 that happened to really be just a
> bit shorter, but since the upgrade the 2 work units that have finished were
> done in 2:58 and 3:00 - normally this system takes about 3:20 per work unit.
>
> Anyone else do the upgrade and seeing anything similar, or am I the first
> guinea pig here?
>
>
>
>

Yep, the GUI's dont show the progress because BOINC v3.19 is using slightly different format in .xml's than the v3.18
ID: 6881 · Report as offensive
WerK

Send message
Joined: 30 Jun 02
Posts: 26
Credit: 221,390
RAC: 0
Czech Republic
Message 6932 - Posted: 12 Jul 2004, 12:12:34 UTC - in response to Message 6736.  

> I updated one of my hosts today to the latest self-compile (boinc 3.20, seti
> 3.10), which seems a little better and a little worse at the same time.
>
> On the better side, it benchmarked with a better whetstone. P4 3.06 1024 RAM
> with whetstone of 917 as compared to 813 on self-compiled version 3.18
> (compiled from the June 24 source code). Also on the better side, while it
> was finishing the 2 work units already in the cache, boincgui was reporting
> the "to completion" times correctly, instead of running it all the way up to
> 50% done in the first minute, then taking a few hours to do the other 50%.
> These work units, of course, ran using seti 3.08 though.
>
> Now that it's gone on to a work unit downloaded after the upgrade and
> processing with seti 3.10, none of the gui add-ons show the progress (they
> read just like predictor units do on linux gui add-ons (such as "to completion
> 0-21467484.....").
>
> Lastly, the work units themselves seem to be processing a slight bit faster.
> It could have just been a fluke and I had 2 that happened to really be just a
> bit shorter, but since the upgrade the 2 work units that have finished were
> done in 2:58 and 3:00 - normally this system takes about 3:20 per work unit.
>
> Anyone else do the upgrade and seeing anything similar, or am I the first
> guinea pig here?
>
>
>
>

Could you please post your app_info.xml file and binaries placement ?? I was trying to get my compiled s@h 3.20 to work but i'm getting some execv error :[ thx
ID: 6932 · Report as offensive
Darren
Volunteer tester
Avatar

Send message
Joined: 2 Jul 99
Posts: 259
Credit: 280,503
RAC: 0
United States
Message 7071 - Posted: 12 Jul 2004, 16:21:39 UTC - in response to Message 6932.  


> Could you please post your app_info.xml file and binaries placement ?? I was
> trying to get my compiled s@h 3.20 to work but i'm getting some execv error :[

Here's the app_info.xml file. As to "binaries placement", I'm gonna need that worded for a little bit dumber person - which binaries and placement in relation to what? If you're simply asking where I put the binaries that were compiled, I put the boinc_client file in my main boinc directory (on my system, /home/darren/boinc/) and the setiathome-3.10.i686-pc-linux-gnu file in the ~/projects/setiathome.berkeley.edu/ subdirectory along with the app_info.xml file.

(app_info)
(app)
(name)setiathome(/name)
(/app)
(file_info)
(name)setiathome-3.10.i686-pc-linux-gnu(/name)
(/file_info)
(app_version)
(app_name)setiathome(/app_name)
(version_num)310(/version_num)
(file_ref)
(file_name)setiathome-3.10.i686-pc-linux-gnu(/file_name) (main_program/)
(/file_ref)
(/app_version)
(/app_info)



ID: 7071 · Report as offensive

Message boards : Number crunching : Version 3.20 on Linux


 
©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.