SETI Beta Application/Optimalization

[trux]

Crunch3r, what is the right CSV server & module name to get the source code for Enhanced? I tried seti_enhanced and setiathome_enhanced that I see listed at alien.ssl.berkeley.edu:/home/cvs/cvsroot, but they are both empty!

[Crunch3r]

Crunch3r, what is the right CSV server & module name to get the source code for Enhanced? I tried seti_enhanced and setiathome_enhanced that I see listed at alien.ssl.berkeley.edu:/home/cvs/cvsroot, but they are both empty!

Don't use the cvs ... it's borken.

Try one of the nightly builds they work http://setiweb.ssl.berkeley.edu/sah/set ... e/nightly/.


P.S. I allready replied to your post earlier today. Seems as if the posting was lost

P.P.S If you're trying to compile this crap have a closer look at the malloc.h implemetation!

regards
Crunch3r

[Sir Ulli]

first Results with my P4 3.2 and HT

with the spezial Version from Crunsh3r, on Seti Enhanced

http://setiweb.ssl.berkeley.edu/beta/re ... ostid=3510

that are 17 Hours for 2 WUs, not Bad i think....

Sir Ulli

[trux]

Yes, it looks not too bad. Just a pity you do not have couple of WU's done with the official Enhanced client for comparison.

[Sir Ulli]

btw

when i look at the Enhanced Forum and the Top crunshers, i found this

CPU type GenuineIntel
Intel(R) Pentium(R) D CPU 3.20GHz

http://setiweb.ssl.berkeley.edu/beta/re ... ostid=2137

and

<core_client_version>4.72</core_client_version>
<stderr_txt>
(seti-enhanced 4.02 with ICC 9.0/fftw by TMR Fri Jul 08 02:04:54 2005)

</stderr_txt>

has TMR also take a look at this


http://setiweb.ssl.berkeley.edu/beta/re ... ostid=2137

Sir Ulli

btw this Host ist Fast

[trux]

Hmm, they seem to be all faster, or am I seeing wrong? Well, they may be overclocked - that's why I told it is a pity you didn't test the official client on the same machine, but finally you can still do it

[Crunch3r]

btw

when i look at the Enhanced Forum and the Top crunshers, i found this

CPU type GenuineIntel
Intel(R) Pentium(R) D CPU 3.20GHz

http://setiweb.ssl.berkeley.edu/beta/re ... ostid=2137

and

<core_client_version>4.72</core_client_version>
<stderr_txt>
(seti-enhanced 4.02 with ICC 9.0/fftw by TMR Fri Jul 08 02:04:54 2005)

</stderr_txt>

has TMR also take a look at this


http://setiweb.ssl.berkeley.edu/beta/re ... ostid=2137

Sir Ulli

btw this Host ist Fast

Hello,

the host you're refering to uses an old optimized "enhanced" app.
Later it was removed my TMR because it had some errors.

BTW.

P.S. Finished my second WU ---> http://setiweb.ssl.berkeley.edu/beta/re ... ostid=3508

[trux]

Yes, 35ks looks much better

[Sir Ulli]

my P4 3.2 with Ht for 2 WUs

http://setiweb.ssl.berkeley.edu/beta/re ... ostid=3510

also not bad

using the Windows Version, what i hear the Linux Version is faster..

Sir Ulli

[Crunch3r]

Yes, 35ks looks much better

The second one finishe in 36k sec.
To bad that they released 5.0 of the enhanced app.

The problem is that the nightly tarballs haven't been updated since 03. Jan.

Have you had a look at the enhanced app or compiled it yet ?

[trux]

Have you had a look at the enhanced app or compiled it yet ?

No, unfortunately, I had to work a bit today, so did not yet start playing with it

[Honza]

To bad that they released 5.0 of the enhanced app.
The problem is that the nightly tarballs haven't been updated since 03. Jan.

So, should we run couple of WUs over weekend using official 5.0 version or should we better wait until optimalized version is out

[Crunch3r]

To bad that they released 5.0 of the enhanced app.
The problem is that the nightly tarballs haven't been updated since 03. Jan.

So, should we run couple of WUs over weekend using official 5.0 version or should we better wait until optimalized version is out

Would be nice if you could run the official 5.0 so that we can get some processing times to compare against the optimized version.

I'm building an optimized version of the 5.0 over the weekend.

[Honza]

Like others reporting on SETI BETA, I have troubles downloading better_banner.jpg (checksum error). It there a workaround available?
Also, I'm getteing outdated setiathome_4.11_windows_intelx86.exe - not the 5.0 version.

CPDN BOINC went down earlier today - both public slab/sulphur and SpinUp

06/01/2006 20:19:04||Network error: couldn't connect to server
06/01/2006 20:19:07|CPDN HadCM3 Spinup|Scheduler list fetch failed: http error
06/01/2006 20:19:07|CPDN HadCM3 Spinup|4 consecutive failures fetching scheduler list - deferring 604800 seconds.

Hmm, lot of troubles with network recetnyl.

Hope to get your optimalized version package with errant .jpg file - this might be good workaround to the problem and can permit do run on official client as well (albeit outdated).
I can test on Pentium D820 or X2 3800+

[Crunch3r]

Like others reporting on SETI BETA, I have troubles downloading better_banner.jpg (checksum error). It there a workaround available?
Also, I'm getteing outdated setiathome_4.11_windows_intelx86.exe - not the 5.0 version.

CPDN BOINC went down earlier today - both public slab/sulphur and SpinUp

06/01/2006 20:19:04||Network error: couldn't connect to server
06/01/2006 20:19:07|CPDN HadCM3 Spinup|Scheduler list fetch failed: http error
06/01/2006 20:19:07|CPDN HadCM3 Spinup|4 consecutive failures fetching scheduler list - deferring 604800 seconds.

Hmm, lot of troubles with network recetnyl.

Hope to get your optimalized version package with errant .jpg file - this might be good workaround to the problem and can permit do run on official client as well (albeit outdated).
I can test on Pentium D820 or X2 3800+

The problem is that i cannot redistribute an optimized version of seti@home enhanced 5.0 before anyone got an official version downlodaed.

I got all versions for linux and windows allready here based on the 5.0 sources. But i don't know if we could use the new version to complete WUs allready started with 4.11 or if the WUs are completely different.

I also asked "Trux" to join us at the "Beta Test" Forum @ the german "Plane3dnow" Board as i'm trying to concentrate all the developers and beta testers there to better keep track of all versions.

I don't want to enable downloads for the optimized enhanced version until we can really asure that it's working properly, but i assume you can understand it.

Anyway it's really strange that the 5.0 version has allready been announced as an official release but nobody was able to download it ... not even if one resets the project.

I hope that you can understand my point, that i cannot relase the clients until they prove to be reliable.

Anyhow.. I invented Trux to join the board/Beta forum because there are issues of claimed credit with the optimized enhanced
app. he speaks perfect german and that will make it easier to coordinate our work.

I'm also trying to get an international "Thread" for all the developers/Beatatesters of SETI&BOINC working over there because it'll ease coordinating things for ... (As Trux is speaking german perfectly is makes it way easier for us all).

But that's not all.. I'll try to get "Harold Naparst" (after he moved from US to Sweden), Ned Slider and a lot of my "beta Testers" Daniel Schalmaa and Al Reust("Pappa") to post on that board.

P.S. It does not mean we wont post here anymore

It just makes things easier to coordinate for me.
Hopefully all people i contacted via e-mail will reply with a positive answer

Anyway. We have a lot to work on and if you feel that running the standard enhanced app doesn't contibute any usefull information, i can asure you that it's really important for us to know!

P.S. If all goes well with my optimized version ... the czech board will be the first to get a link to my release!

Once again i have to ask for you patience.

I hope you know that just compiling the source doesn't make the app. work faster and it uses a lot of testing before it can be released so that i can't do it now before it works and i'm concentating on a 64 bit version for linux that can be ported to Win64.
(But that's not another story and not that important)

regards
Crunch3r