+- +-
Say hello if visiting :) by Gecko
11 Jan 2023, 07:43:05 pm

Seti is down again by Mike
09 Aug 2017, 10:02:44 am

Some considerations regarding OpenCL MultiBeam app tuning from algorithm view by Raistmer
11 Dec 2016, 06:30:56 am

Loading APU to the limit: performance considerations by Mike
05 Nov 2016, 06:49:26 am

Better sleep on Windows - new round by Raistmer
26 Aug 2016, 02:02:31 pm

Author Topic: V10/11 of modified SETI MB CUDA + opt AP package for full multi-GPU+CPU use  (Read 233385 times)

Offline Raistmer

  • Working Code Wizard
  • Volunteer Developer
  • Knight who says 'Ni!'
  • *****
  • Posts: 14349
See first post, it stated as clearly as I could formulate that.
Here I better describe why it should be so.
In team mod BOINC doesn't know at all that some of apps it controls using GPU/CUDA.
It manages all applications as CPU ones. That is, it will run only ncpus copies of AK_v8 for example. But you need to take all cores and all GPUs busy, that's why you need number of started by BOINC apps to be NUMBER_OF_CORES + NUMBER_OF_GPUs in V10 packs.

Offline Jason G

  • Construction Fraggle
  • Knight who says 'Ni!'
  • *****
  • Posts: 8980
So.... not advised to run with current Astropulse only work climate then, lest you get NumGPUs over subscription... right?

Kashra

  • Guest
I just tried to install V10a x64 build on my XP x64 but it did not work.
I alway get a computation error.

Here is what I did:

Installed Boinc 6.4.5
Let Boinc fetch standard app + WUs
stop Boinc Service
place the files from the V10 Package in the Project directory and setup the cc_config.xml for 5 (4CPU + 1 GPU) and placed it in the main Boinc "user" directory.
start Boinc service
Both APPs Astro+Seti produce only computation errors all the time but I don´t know why.

Any idea what might cause the error?
The stock version runs without errors.

My system specs:

Intel Q9550 CPU
1x Nvidia GTX280 with 182.06
8 GB DDR-2 Ecc Ram
Windows XP x64 SP2
Boinc 6.4.5 service installation

Offline Richard Haselgrove

  • Messenger Pigeon
  • Knight who says 'Ni!'
  • *****
  • Posts: 2819
It might be easier if you told us what the error messages actually said.

The strong, silent type ;D There's a Kashra at SETI who runs host 4493623 - Q9450 and GTX280.

The error message (consistently) is

"Es gibt keine untergeordneten Prozesse, auf die gewartet werden muss. (0x80) - exit code 128 (0x80)"
« Last Edit: 28 Feb 2009, 07:14:12 am by Richard Haselgrove »

Offline Jason G

  • Construction Fraggle
  • Knight who says 'Ni!'
  • *****
  • Posts: 8980
"Es gibt keine untergeordneten Prozesse, auf die gewartet werden muss. (0x80) - exit code 128 (0x80)"
  awww, without knowing what that is saying, it still 'sounds' rather coarse. [Hopefully Urs or Crunch3r might rock up, and make more sense of it]
« Last Edit: 28 Feb 2009, 07:22:28 am by Jason G »

Offline Richard Haselgrove

  • Messenger Pigeon
  • Knight who says 'Ni!'
  • *****
  • Posts: 2819
Google thinks "There are no child processes to wait to be".

Offline Jason G

  • Construction Fraggle
  • Knight who says 'Ni!'
  • *****
  • Posts: 8980
How about "Could not start Child Process?" .  Maybe there is some executable from the package that was not put in the project folder?

Kashra

  • Guest
I finally figured it out now.
The problem seems to be the protected mode installation of Boinc (Boinc runs as service).
After reconfiguring Boinc not to run in protected mode the V10 App runs just fine. I know protected mode will not work with Vista, but xp 64 should do fine - but it does not at least with V10.

I wonder why it works with the stock version in protected mode but not with V10.

Thanks for getting back to me this fast.  :) The error message on the stats site "Could not start Child Process" was the clue.


Offline Jason G

  • Construction Fraggle
  • Knight who says 'Ni!'
  • *****
  • Posts: 8980
Ahhh, The Old "Cuda does not like Boinc as Service/Protected Application in Vista" trick eh?(Maybe applies to x64?)  Will remember that for the next customer!. Thanks for checking in and persevering  :D

SoNic

  • Guest
Thank you for the new release, it is working great on my system.
I have one complaint thou. Throwing out the VLAR WU... It is a shame, I was just looking to the Boinc client and I saw 12 of them beeing rejected one after another. I don't think it is fair to the other wingman.
Is there a way that this can be solved without autokill? Like force them to the CPU instead of GPU? What happens if there are not killed - under V10 app?

Offline Urs Echternacht

  • Volunteer Developer
  • Knight who says 'Ni!'
  • *****
  • Posts: 4121
  • ++
Google was close on this one - should be :
"There are no child processes to wait for". ;)
_\|/_
U r s

Offline Jason G

  • Construction Fraggle
  • Knight who says 'Ni!'
  • *****
  • Posts: 8980
LoL, Thanks Urs'   I have no Children yet either  :(

Offline Urs Echternacht

  • Volunteer Developer
  • Knight who says 'Ni!'
  • *****
  • Posts: 4121
  • ++
Just found this (possible/temporary) solution in this forum post.

Short : Try that GLtest does NOT run. If that is the case, add the file MSCorEE.dll into the (windows)\\system32 directory.

Anyone feels to try out this ? Seems that this is a pure BOINC problem.



edit: ooppsee : "There are no subprocesses to wait for". Does it make sense now.
_\|/_
U r s

Offline Jason G

  • Construction Fraggle
  • Knight who says 'Ni!'
  • *****
  • Posts: 8980
...does NOT run. If that is the case, add the file MSCorEE.dll into the (windows)\\system32 directory.
...

This would make perfect sense to me if so (can't test it though, currently no active Vista or x64 install),  It means that either nVidia or Boinc failed to distribute the correct dependencies within their respective installers  (either or both, depending on how you look at it)

[Edit: Incidentally, GLTest.exe runs fine on my Windows XP w/SP3 + eVGA 9600GSO, nVidia Driver version 181.22 ]
« Last Edit: 28 Feb 2009, 10:13:57 am by Jason G »

P51 Mustang

  • Guest
My computer has been throwing a lot of computation errors since I started using V10a x64.  I went from V8 to V10a to get the new AP units.  It is a Quad core running Vista 64 bit with driver 181.22 with a single GeForce 8500 GT.

Some of the errors are the VLAR kill, but most of them are not.  Here is a link to a workunit that is throwing the computation error.

http://setiathome.berkeley.edu/result.php?resultid=1173667098
try to revert to old DLL versions (from older packs or stock).
Lookd like driver/CUDA runtime problem... If with same DLL/driver set stock app doesn't show such errors, report again please.


Well, everything looked to run OK with the older dll files.  Redownloaded the dlls for V10 and it looks like some go through ok while others fail.  Here are links to the wu's.

http://setiathome.berkeley.edu/result.php?resultid=1174424840  completed successfully
http://setiathome.berkeley.edu/result.php?resultid=1174852750  computation error

I've just installed the latest drivers from Nvidia 182.06 and will see how that works out.

 

Welcome, Guest.
Please login or register.
 
 
 
Forgot your password?
Members
Total Members: 97
Latest: ToeBee
New This Month: 0
New This Week: 0
New Today: 0
Stats
Total Posts: 59559
Total Topics: 1672
Most Online Today: 18
Most Online Ever: 983
(20 Jan 2020, 03:17:55 pm)
Users Online
Members: 0
Guests: 8
Total: 8
Powered by EzPortal