Seti@Home optimized science apps and information

Optimized Seti@Home apps => Linux => Topic started by: michael37 on 30 Mar 2007, 12:18:59 am

Title: FIXED - Still problems with statified initial public release of Linux Core 2 Duo
Post by: michael37 on 30 Mar 2007, 12:18:59 am
Running standalone version with work_unit.sah in directory. 

Side question: what the heck does SSE3-C2 find for my processor -- see wierd CPUs and cache?  I will include stderr by SSE3-P4 application which looks normal.

Code: [Select]
[Files-to-install]# ./KWSN-R2.2B-SSE3-C2
Aborted
[Files-to-install]# echo $?
134
[Files-to-install]# cat stderr.txt
Can't open init data file - running in standalone mode
Can't open init data file - running in standalone mode
Optimized SETI@Home Enhanced application
Optimizers: Ben Herndon, Josef Segur, Alex Kan, Simon Zadra
Linux port: Crunch3r, Hans Dorn, Simon Zadra
   Version: Linux 32-bit based on S@H V5.15  'Noo? No - Ni!'
  Revision: R-2.2B|xT|FFT:IPP_SSSE3|Ben-Joe
     CPUID: Intel(R) Xeon(R) CPU            5130  @ 2.00GHz
      CPUs: 0, cores: 1,  cache: L1=2K,  L2=0K, L3=0K
  Features: MMX  SSE  SSE2  SSE3  SSSE3
Title: Re: Still problems with statified initial public release of Linux Core 2 Duo
Post by: michael37 on 30 Mar 2007, 12:19:36 am
stderr.txt from SSE3-P4 application.

Code: [Select]
Can't open init data file - running in standalone mode
Can't open init data file - running in standalone mode
Optimized SETI@Home Enhanced application
Optimizers: Ben Herndon, Josef Segur, Alex Kan, Simon Zadra
Linux port: Crunch3r, Hans Dorn, Simon Zadra
   Version: Linux 32-bit based on S@H V5.15  'Noo? No - Ni!'
  Revision: R-2.2B|xP|FFT:IPP_SSE3|Ben-Joe
     CPUID: Intel(R) Xeon(R) CPU            5130  @ 2.00GHz
      CPUs: 1, cores: 1,  cache: L1=32K,  L2=4096K, L3=4096K
  Features: MMX  SSE  SSE2  SSE3   
     speed: 1995 MHz  -- read MB/s: L1=7597, L2=7002, L3=6361, RAM=2720

Work Unit Info
True angle range:  0.474242
Title: Re: Still problems with statified initial public release of Linux Core 2 Duo
Post by: Crunch3r on 30 Mar 2007, 05:01:06 am
stderr.txt from SSE3-P4 application.

Code: [Select]
Can't open init data file - running in standalone mode
Can't open init data file - running in standalone mode
Optimized SETI@Home Enhanced application
Optimizers: Ben Herndon, Josef Segur, Alex Kan, Simon Zadra
Linux port: Crunch3r, Hans Dorn, Simon Zadra
   Version: Linux 32-bit based on S@H V5.15  'Noo? No - Ni!'
  Revision: R-2.2B|xP|FFT:IPP_SSE3|Ben-Joe
     CPUID: Intel(R) Xeon(R) CPU            5130  @ 2.00GHz
      CPUs: 1, cores: 1,  cache: L1=32K,  L2=4096K, L3=4096K
  Features: MMX  SSE  SSE2  SSE3   
     speed: 1995 MHz  -- read MB/s: L1=7597, L2=7002, L3=6361, RAM=2720

Work Unit Info
True angle range:  0.474242

Hi,

I just wouldn't bother about the output regarding the cache size.

While porting the code i ran into lot's of trouble regarding Bens CPUID which didn't work at all so i had to replace it with something "home brewed".

Same goes for the cache size, while compiling the source on gentoo, the app would have shown "0K" for all caches... Same source compiled on SuSE 9.2 and it works.

So this part seems to depended on the distro you run, too.



Title: Re: Still problems with statified initial public release of Linux Core 2 Duo
Post by: nastasache on 30 Mar 2007, 03:13:34 pm
Hi.
I use an E6600 C2D running under Gentoo 2006.1.

No problem to date with old SSE2 apps for P4 (setiathome-5.15-kwsn-1.3-sse2-p4).

Today I tried to switch to SS3 for my cpu (KWSN-R2.2B-SSE3-C2) but I have received a lot of computation error  with "process got signal 6" - see: http://setiathome.berkeley.edu/result.php?resultid=508731854

Same like in michael37 case, using SSE3-P4 application looks working normally. I tested KWSN-R2.2B-SSE3-P4 instead of KWSN-R2.2B-SSE3-C2, and P4 looks ok.

Still, where may the problem more exactly? Or, I missing something else important?

I'am under Gentoo x86_64 using same sys/kernel/libraries  for all above versions.

My machine:
# uname -a
Linux oblio 2.6.20-gentoo-r3 #1 SMP Sun Mar 18 16:32:47 EET 2007 x86_64 Intel(R) Core(TM)2 CPU  6600  @ 2.40GHz GenuineIntel GNU/Linux

For the moment, i switched to setiathome-5.15.i686-pc-linux-gnu  from http://setiathome.berkeley.edu/forum_thread.php?id=38058 and looks working good, but I am not sure that last is P4 or C2 optimized. Can you confirm it is P4 or C2? What is better to use if KWSN-R2.2B-SSE3-C2 I can't?

Thanks,
Iulian


 
Title: Re: Still problems with statified initial public release of Linux Core 2 Duo
Post by: fubared on 01 Apr 2007, 05:42:31 am
I have a Core2 on a gentoo x64 based box.

I also get a signal 6 error on the c2d app.

I have no idea why its erroring but I'm guessing your working on a fix I presume?
Title: Re: Still problems with statified initial public release of Linux Core 2 Duo
Post by: michael37 on 01 Apr 2007, 08:34:23 am
I have a Core2 on a gentoo x64 based box.

I also get a signal 6 error on the c2d app.

I have no idea why its erroring but I'm guessing your working on a fix I presume?

So far, it doesn't look like the complete C2 application will work on Linux on most distributions.  It has to do with glibc libraries.   Current recommendation is to use the SSE3-P4 application.
Title: Re: Still problems with statified initial public release of Linux Core 2 Duo
Post by: Natronomonas on 02 Apr 2007, 04:52:50 am
Not only have I had a 'signal 6' exit, I've also got a 'maximum CPU time exceeded' error (Ubuntu 6.10).

If only I'd checked back here recently; just lost about 3 days crunching time on this computer (not on the network often, I just let it sit there). Oh well, that's the way the cookie crumbles.

Anyone know how long it takes to get the cpu time exceeded error?

edit: I have glibc 2.4 installed, if that helps any.

edit#2: The generic SSE 2.2B client is working fine brilliantly on exactly the same version of Ubuntu (AthlonXP machine).
So it does seem (as others have said) to be limited to the C2D app. If there is anything the devs require please do not hesitate to ask - a lowly knave like myself is only too happy to help : )
Title: Re: Still problems with statified initial public release of Linux Core 2 Duo
Post by: Josef W. Segur on 02 Apr 2007, 02:17:31 pm
...
Anyone know how long it takes to get the cpu time exceeded error?
...

For S@H WUs, it's the raw completion estimate x10, which works out to more than normal crunch time x20 on most hosts.
                                                                                     Joe