Posts by mr.kjellen

1) Message boards : Number crunching : What's your oldest unvalidated WU ? (Message 1053212)
Posted 5 Dec 2010 by Profile mr.kjellen
Post:
I'm kinda close to the 13th...May 26th. No cigar.
2) Message boards : Number crunching : Win7 x64 + GTX295 + PhysX (Message 946877)
Posted 13 Nov 2009 by Profile mr.kjellen
Post:
These 295's are a bit tricky to say the least...
As of driver set 191.07 on an x64 vista machine I am required to use SLI for both cores of the card to be picked up by boinc, at least on the single PCB version.
Also the flag use all gpus needs to be ticked in the cc_config file.

But when I read of other users experiences the requirements for a working setup seem to vary a lot.. or mabye there are just several different ways to make it work...

/Anton
3) Message boards : Number crunching : Best Graphics Setup? (Message 938356)
Posted 8 Oct 2009 by Profile mr.kjellen
Post:


I don't think anybody on Seti has been able to show an improvement in crunching performance with the addition of SSDs.......
Sure, they may help overall system performance, and could appear to make the rig faster by speeding up booting and running other programs....but nobody has documented a crunching boost.

Correct me if I am wrong.


Hmm, I came to think about the issue with a large cache causing massive boinc slowdowns...did that ever get traced back to HDD writes? Could an SSD help out in a case with say a 4000 unit cache?
4) Message boards : Number crunching : nVIDIA driver 191.07 released. (Message 938289)
Posted 8 Oct 2009 by Profile mr.kjellen
Post:
My i7 with two different GPUs (8800GTS & 9800+) just froze up with 191.07. I went back to driver 190.38, no such issues since.

/Anton
5) Questions and Answers : GPU applications : Reported clockrate value in stderr_txt reduced over time (Message 929886)
Posted 31 Aug 2009 by Profile mr.kjellen
Post:
update,
It seems to be the thermal throttle after all. after reviewing the logs from gpuz I can see that the cards do throttle every once in a while, only in short spurs lasting perhaps 10-15 seconds.

So my guess is boinc picks up on that and reports it, eventhough the downclock isn't permanent.

/Anton
6) Questions and Answers : GPU applications : Reported clockrate value in stderr_txt reduced over time (Message 928748)
Posted 26 Aug 2009 by Profile mr.kjellen
Post:
Hi, I'm running this rig It has 2 GTX295s that are oc'd to 665/1100 (from i believe 575/1000).

The thing is the reported clockrate value on card 2 and 4 are reduced over time. After a reboot they report correctly again. GPU-z, Riva tuner and Nv Control panel always report the correct (665) value.

The theory I'm going with now is that this is some sort of thermal throttle kicking in, though I have never seen a card go above 85C.

Example from a task stderr:

setiathome_CUDA: Found 4 CUDA device(s):
Device 1 : GeForce GTX 295
totalGlobalMem = 939261952
sharedMemPerBlock = 16384
regsPerBlock = 16384
warpSize = 32
memPitch = 262144
maxThreadsPerBlock = 512
clockRate = 1440000
totalConstMem = 65536
major = 1
minor = 3
textureAlignment = 256
deviceOverlap = 1
multiProcessorCount = 30
Device 2 : GeForce GTX 295
totalGlobalMem = 939196416
sharedMemPerBlock = 16384
regsPerBlock = 16384
warpSize = 32
memPitch = 262144
maxThreadsPerBlock = 512
clockRate = 601714
totalConstMem = 65536
major = 1
minor = 3
textureAlignment = 256
deviceOverlap = 1
multiProcessorCount = 30
Device 3 : GeForce GTX 295
totalGlobalMem = 939261952
sharedMemPerBlock = 16384
regsPerBlock = 16384
warpSize = 32
memPitch = 262144
maxThreadsPerBlock = 512
clockRate = 1440000
totalConstMem = 65536
major = 1
minor = 3
textureAlignment = 256
deviceOverlap = 1
multiProcessorCount = 30
Device 4 : GeForce GTX 295
totalGlobalMem = 939261952
sharedMemPerBlock = 16384
regsPerBlock = 16384
warpSize = 32
memPitch = 262144
maxThreadsPerBlock = 512
clockRate = 799200
totalConstMem = 65536
major = 1
minor = 3
textureAlignment = 256
deviceOverlap = 1
multiProcessorCount = 30
setiathome_CUDA: CUDA Device 4 specified, checking...
Device 4: GeForce GTX 295 is okay

They run Raistmers V12 app.

Should I trust the Boinc report or GPUz/Riva tuner/Nvida Control Panel? Is this a problem at all?

/Anton
7) Message boards : Number crunching : *CLOSED SETI/BOINC Milestones (tm) XVIII /CLOSED* (Message 927674)
Posted 21 Aug 2009 by Profile mr.kjellen
Post:
Seems I'll venture north of
20 000 000 this weekend! Woot!

/Anton
8) Message boards : Number crunching : Q35 Motherboard OCing (Message 871645)
Posted 3 Mar 2009 by Profile mr.kjellen
Post:


The mod page says the BSEL mod will not work on GIGABYTE motherboards.

snip
Martin


Too bad...What can I say, sorry but I'm an Asus guy. When it works the BSEL mod is great. Easy, non destructable and reversible.

9) Message boards : Number crunching : Q35 Motherboard OCing (Message 871429)
Posted 2 Mar 2009 by Profile mr.kjellen
Post:
As long as you have a motherboard and memory that can handle an FSB of 1600, you should be golden. Do this and your Q9450 should act as a locked XQ9770 (that is 1600FSB and 3.2GHz).

I looked at the specs for your motherboard at the gigabyte site, and it seems the board has 800-1333 FSB support. It MAY still support FSB1600, but it's not certain to work...
10) Message boards : Number crunching : Q35 Motherboard OCing (Message 871406)
Posted 2 Mar 2009 by Profile mr.kjellen
Post:
It involves covering one or two pads underneath the CPU with a small piece of electrical tape, specifically the pads that tell the motherboard at witch FSB the cpu is supposed to operate.


Something along these lines.

Edit:link
11) Message boards : Number crunching : Q35 Motherboard OCing (Message 871347)
Posted 2 Mar 2009 by Profile mr.kjellen
Post:
Hey Raistmer, just throwing an idea out there for you;
If the mobo itself is not OCable, how about a BSEL mod then to fool it to run at a higher FSB? (asuming your CPU isn't already at 1600)

/Anton
12) Message boards : Number crunching : Preferences: No Astropulse (Message 869621)
Posted 26 Feb 2009 by Profile mr.kjellen
Post:
Yup,
3 applications at the moment:
- 'setiathome_enhanced' (MultiBeam 6.03 cpu &/or 6.08 cuda)
- 'astropulse' (astropulse 5.00 tasks only, all resends these days)
- 'Other Applications' (astropulse_v5 5.03 tasks & both other kinds of work too)

So you'll only get AP5.03 if you tick that third box in preferences. To get a 'semi-balanced' mix of all three, you'd need to untick the first two boxes, and tick 'Accept work from other applications'

Not sure if, when they get around to it, astropulse_v5(5.03) will get its own tick box, or inherit the old 'Astropulse' one when that is shut down completely.

Jason


Astropulse V5 has its own tick box now...
13) Message boards : Number crunching : V10 of modified SETI MB CUDA + opt AP package for full multi-GPU+CPU use (Message 869263)
Posted 25 Feb 2009 by Profile mr.kjellen
Post:
I keep getting these exit with zero status but without finished file since the switch over to V10. The message displays multiple times for the same workunit.

qutote from BOINC:
02/25/09 09:21:40|SETI@home|Restarting task 03ja09ad.7389.5389.6.8.96_1 using setiathome_enhanced version 608
02/25/09 09:22:14|SETI@home|Task 03ja09ad.7389.5389.6.8.26_1 exited with zero status but no 'finished' file
02/25/09 09:22:14|SETI@home|If this happens repeatedly you may need to reset the project.
02/25/09 09:22:14|SETI@home|Restarting task 03ja09ad.7389.5389.6.8.26_1 using setiathome_enhanced version 608
02/25/09 09:22:47|SETI@home|Task 03ja09ad.7389.5389.6.8.96_1 exited with zero status but no 'finished' file
02/25/09 09:22:47|SETI@home|If this happens repeatedly you may need to reset the project.
02/25/09 09:22:47|SETI@home|Restarting task 03ja09ad.7389.5389.6.8.96_1 using setiathome_enhanced version 608
02/25/09 09:23:26|SETI@home|Task 03ja09ad.7389.5389.6.8.26_1 exited with zero status but no 'finished' file
02/25/09 09:23:26|SETI@home|If this happens repeatedly you may need to reset the project.
02/25/09 09:23:26|SETI@home|Restarting task 03ja09ad.7389.5389.6.8.26_1 using setiathome_enhanced version 608
02/25/09 09:23:57|SETI@home|Task 03ja09ad.7389.5389.6.8.96_1 exited with zero status but no 'finished' file
02/25/09 09:23:57|SETI@home|If this happens repeatedly you may need to reset the project.
02/25/09 09:23:57|SETI@home|Restarting task 03ja09ad.7389.5389.6.8.96_1 using setiathome_enhanced version 608
02/25/09 09:24:41|SETI@home|Task 03ja09ad.7389.5389.6.8.26_1 exited with zero status but no 'finished' file
02/25/09 09:24:41|SETI@home|If this happens repeatedly you may need to reset the project.
02/25/09 09:24:41|SETI@home|Restarting task 03ja09ad.7389.5389.6.8.26_1 using setiathome_enhanced version 608
02/25/09 09:24:57|SETI@home|Task 03ja09ad.7389.5389.6.8.96_1 exited with zero status but no 'finished' file
02/25/09 09:24:57|SETI@home|If this happens repeatedly you may need to reset the project.
02/25/09 09:24:57|SETI@home|Restarting task 03ja09ad.7389.5389.6.8.96_1 using setiathome_enhanced version 608

Has this something to do with the (your) app, or is it just a coinsidence?
The computer in question is hostid=4242105

/Anton

Edit: Went back to V9 Multi_GPU and no more of these messages. Now that there are AP's out again, this will work just fine.
14) Message boards : Number crunching : V10 of modified SETI MB CUDA + opt AP package for full multi-GPU+CPU use (Message 869254)
Posted 25 Feb 2009 by Profile mr.kjellen
Post:
Seems to run fine on my x64 vista machine.

Question:
This host is a core i7 with a GTX295, so I set Number of CPUs to 10 and number of GPUs to 2, right?

Now the real question, if i specify in preferences for boinc to use 99% of the processors (thereby disabling one cpu core, I like to keep one core free for general computing...), will it still work both cores of the GPU?

I've been trying to keep an eye on the processes runing, and this seems to be the case. Can I have confirmation of that please.
/Anton
15) Message boards : Number crunching : Preferences: No Astropulse (Message 869053)
Posted 24 Feb 2009 by Profile mr.kjellen
Post:
There's certainly still something incomplete about the transition from Astropulse to Astropulse_v5

I got this overnight:

24/02/2009 04:49:25|SETI@home|Message from server: No work can be sent for the applications you have selected
24/02/2009 04:49:25|SETI@home|Message from server: No work is available for SETI@home Enhanced
24/02/2009 04:49:25|SETI@home|Message from server: No work is available for Astropulse
24/02/2009 04:49:25|SETI@home|Message from server: You have selected to receive work from other applications if no work is available for the applications you selected
24/02/2009 04:49:25|SETI@home|Message from server: Sending work from other applications
24/02/2009 04:49:27|SETI@home|Started download of ap_08ja09aa_B0_P0_00136_20090223_25785.wu
24/02/2009 04:49:27|SETI@home|Started download of ap_08ja09aa_B0_P1_00002_20090223_16935.wu
24/02/2009 04:53:57|SETI@home|Finished download of ap_08ja09aa_B0_P1_00002_20090223_16935.wu
24/02/2009 04:54:06|SETI@home|Finished download of ap_08ja09aa_B0_P0_00136_20090223_25785.wu

The work I got was Astropulse_v5 (503).


I made the same observation. The checkbox on the preferences page must still refer to AP 5.0, not to V5...It was a similar way with the server status, it used to only show AP 5.0 as 'ready to send'.
16) Questions and Answers : Windows : Frequent 'waiting for client' popups (Message 863626)
Posted 8 Feb 2009 by Profile mr.kjellen
Post:
Sounds like it's just waiting for the client to start up.


That would have been my guess as well if it wasn't for all the popups that show when the manager and client is already running and have been for several hours.
17) Questions and Answers : Windows : Frequent 'waiting for client' popups (Message 863613)
Posted 8 Feb 2009 by Profile mr.kjellen
Post:
It's a local install on vista that runs cuda as well, so the manager starts the boinc client. It is running at the time the popups appear. They show every now and then. It does not seem to matter if I'm actually doing anyting in the manager, changing tabs and such.


The message is:
"Connecting to the client. Please wait..."
This goes away by it self after (I assume) connection is made.

I'll grab a screenie when I can. The machine is at work, wich I'm not....
18) Questions and Answers : Windows : Frequent 'waiting for client' popups (Message 863586)
Posted 8 Feb 2009 by Profile mr.kjellen
Post:
hmm,
I'm getting frequent popups telling me that the BOINC manager is waiting for the client. Havn't seen this before.
I have been doing lots of changes recently with the CUDA release and all.

boinc 6.6.4 (6.5 does this also) x64, on a core i7 with GTX295.

Is this just the nature of the beta state of the software?
19) Message boards : Number crunching : BOINC 6.6.4 (Message 863562)
Posted 8 Feb 2009 by Profile mr.kjellen
Post:
Just confirming Richards 'trick' does work on windows (x64) also. Exchanged the 6.6.4 manager with the 6.5 one and it works.
20) Message boards : Number crunching : Running silent? (Message 862668)
Posted 6 Feb 2009 by Profile mr.kjellen
Post:
The easiest way is to do a service install of BOINC. The unofficial BOINC wiki contains how-to guides.

Please note though, you do have to have the permission of the owner to run BOINC... (just making sure :) )

/Anton


Next 20


 
©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.