| Author |
Message |
|---|
|
|
|
I run dual gtx560 ti's with a fx-8120 and every day i get about 2-4 random compute errors in which they all say the environment is wrong here is one.
<core_client_version>6.12.34</core_client_version>
<![CDATA[
<message>
The environment is incorrect. (0xa) - exit code 10 (0xa)
</message>
<stderr_txt>
GeneferCUDA-boinc 1.06 (CUDA3.2) based on GeneferCUDA 1.049 and Genefer 2.2.1
Copyright (C) 2001-2003, Yves Gallot (v1.3)
Copyright (C) 2009-2011, Mark Rodenkirch, David Underbakke (v2.2.1)
Copyright (C) 2010-2012, Shoichiro Yamada (CUDA)
Portions of this software written by Michael Goetz 2011-2012 (BOINC)
A program for finding large probable generalized Fermat primes.
Command line: projects/www.primegrid.com/primegrid_genefer_1.06_windows_intelx86__cuda32_13.exe -boinc -q hidden --device 0
Priority change succeeded.
GPU=GeForce GTX 560 Ti
Global memory=1073741824 Shared memory/block=49152 Registers/block=32768 Warp size=32
Max threads/block=1024
Max thread dim=1024 1024 64
Max grid=65535 65535 65535
CC=2.1
Clock=1850 MHz
# of MP=8
No project preference specified; using SHIFT=7
maxErr during b^N initialization = 0.0000 (0.190 seconds).
Testing b^262144+1...
maxErr exceeded for 734444^262144+1, 0.5000 > 0.4500
10:42:33 (3884): called boinc_finish
</stderr_txt>
]]>
Any ideas i have tried stock clocks over clocked does not seem to make a difference. Or if anyone else has seen these errors consistantly
Either fails right away or after about 2000 seconds so there is no pattern i can find. here is the link to one of my compute erros
http://www.primegrid.com/result.php?resultid=351274101 |
|
|
 Michael Goetz Volunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
                           
|
maxErr exceeded for 734444^262144+1, 0.5000 > 0.4500
That's what error code 10 means. Ignore what BOINC says about the environment.
This is caused by a combination of clock rate and/or temperature.
Best advice is to make the cards as cool as possible (blow the dust out with compressed air and set the fan to run faster), slow down the shader clocks to stock, and slow the memory clock to stock (or even lower).
This will become even more important once we start doing the real work and you'll need to run without error for over a week for each WU.
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
10-4 They run about 70-72 C clocked to 1900-2000 on the shader. stock on these is 900/1800/2004. I shall run them down to 850 and adj my fan profile and see what happens. Thanks |
|
|
Michael GoetzVolunteer moderator
Project administrator
Send message
Joined: 21 Jan 10
Posts: 13513
ID: 53948
Credit: 237,712,514
RAC: 0
|
10-4 They run about 70-72 C clocked to 1900-2000 on the shader. stock on these is 900/1800/2004. I shall run them down to 850 and adj my fan profile and see what happens. Thanks
Actually, stock for a 560 Ti is 1645. The 1800 is probably factory overclocking. It is sometimes necessary to go to the real stock clock speed regardless of the factory setting.
http://www.nvidia.com/object/product-geforce-gtx-560ti-us.html
____________
My lucky number is 75898524288+1 |
|
|
|
|
|
You are correct 822 is stock. i guess i should have specified the factory overclock. but i have discovered that the only one that is actually erroring out is my 1GB (which was a refurb from pny i might add) i have clocked it to 822 down from 900 (factory clock). My EVGA i bought retail runs stable up to 1025/2050/2400. but like you said when the real work begins i dont wanna loose 3-5 days of work just to save a few hours of crunch time... I will keep a close eye on WU's to make sure im not comp error'in them out..
Now if we could just figure out how to make the electricity bill go down :)
|
|
|
|
|
|
Run OCCT test and see if any, even one error occur in at least one hour test.
OCCT can be downloaded from http://www.ocbase.com/perestroika_en/index.php?Download
Even one error mean your card is not stable. I have 560 Ti, also factory OC and wasnot stable until I raise voltage. Maybe it is your case also?
____________
92*10^1439761-1 REPDIGIT PRIME :) :) :)
314187728^131072+1 GENERALIZED FERMAT
31*332^367560+1 CRUS PRIME
Proud member of team Aggie The Pew. Go Aggie! |
|
|
|
|
|
to make the PNY (Refurb) stable i have to crank the voltage to around 1.10
ori can just run it at 1.050 at 850 and loose 5 min per work unit....im debating which way to go....:) i am gonna chk out your program you suggested....Thanks. |
|
|
|
|
|
My run time was around 3515 seconds per WU using Genefer cuda 1.06 block size 7 ( now it is default)
I have 560 Ti OC to 900 with 1 GB DDR5
You have 2 GB version is it "new" 560 TI with 448 cores or just one revision of "ordinary" 560 Ti?
____________
92*10^1439761-1 REPDIGIT PRIME :) :) :)
314187728^131072+1 GENERALIZED FERMAT
31*332^367560+1 CRUS PRIME
Proud member of team Aggie The Pew. Go Aggie! |
|
|
|
|
Now if we could just figure out how to make the electricity bill go down :)
Buy a 80 Plus gold or platinum PSU. You'd be surprised at how much power you save, even over a bronze one. And very little heat from it means less on A/C in the summer so even less electricity use. And the PSU will last much longer. Lowering clocks on your GPU instead of raising voltage or a compromise of sorts will help a lot too. I know you said that in jest but this is still good info for anyone reading.
NM*
____________
Largest Primes to Date:
As Double Checker: SR5 109208*5^1816285+1 Dgts-1,269,534
As Initial Finder: SR5 243944*5^1258576-1 Dgts-879,713
 |
|
|
|
|
|
Yea i got 2-800watt bronze PSU's right now with a 400 watt strapped in line with my 8-core and 2 560 ti... system a 800 watt will do it but it run close to 95 % all the time so i put a 400 on the 560ti and harddrive just to take some of the strain off my 800 ...seems alot happier now. |
|
|
