Gauntlet 2009 Joiners Thread!

shadowscotland said:
Everones graphs on the previous page are looking very health and with gains across the board.
Turning into a real race :D (thanks for the running updates simon)
Click to expand...

No problem,

Was interested to see how we were doing and it looks OK, we need to get a bit more on a daily basis to get closer to OCForums but I am optimistic esp with news that TT has got his 250 arriving soon.

I have tweaked the cards in my main machine upto 657/1457/1100 which gives a good boost. If you believe Boinc's estimates they go from a stock 106Gflops to 125Gflops per card. They seemed stable overnight and I will check again when I finish work.

Looking forward to tonights numbers.

Cheers, Simon.
 
Last edited:
Right, I've got me 250, stuck it in and installed the 190.62 drivers.

Now what? :p

I've downloaded the following files:

  • MB_6.08_CUDA_V12_VLARKill_FPLim2048_test.exe
  • cudart.dll
  • cufft.dll

The .dlls came from the Cuda2.3rt_fft_dlls.7z file on the Lunatics site, if that's the right one to be using. I assume I place the above files in the Data\Projects\setiathome.berkeley.edu directory, but I also assume I need to alter app_info.xml to point to them - and I don't know what I need to put in there :o. Any pointers?
 
Hi,

Good news :)

Have a look at page 1 of this thread about two thirds of the way down there are a couple of posts about the install, it shows a link to the Lunatics Unified Installer and a couple of pics also.

I am still at work but will check back later

Cheers, Simon.
 
TheTross said:
Right, I've got me 250, stuck it in and installed the 190.62 drivers.

Now what? :p

I've downloaded the following files:

  • MB_6.08_CUDA_V12_VLARKill_FPLim2048_test.exe
  • cudart.dll
  • cufft.dll
The .dlls came from the Cuda2.3rt_fft_dlls.7z file on the Lunatics site, if that's the right one to be using. I assume I place the above files in the Data\Projects\setiathome.berkeley.edu directory, but I also assume I need to alter app_info.xml to point to them - and I don't know what I need to put in there :o. Any pointers?
Click to expand...

TT,

Backup your BOINC folder first (just in case!)

Use the Lunatics v0.2 installer. It will create the app_info.xml file for you.

Select the apps according to your CPU capabilities (SSSE3 is usually better than SSE4.1). You will need to copy the v2.3 CUDA dlls over the v2.2 versions from the installer. After that you should be good to go.
 
Last edited:
TT - you should just be able to use CUDA 2.3 from the nvidia website as those dll's were only released on lunatics as they offered a big improvement over the then available CUDA 2.2.
Start boinc, attach to project and then use the lunatics unified installer as it will install the correct apps for you. If you are unsure what to use cpu-z is your friend!
 
Well, it's up and running :D. Surprisingly silent too, as the fan's running at an inaudible 35% with the core hovering around 65°C.

Have just put the v2.3 dlls in now; they seem to be a fair bit faster than the v2.2 ones.

Thanks everyone :)


Edit: Daft question time perhaps, but this card has two 6-pin power connectors on it whereas all the other 250s I've seen only have one :confused:. I've used both (both directly with separate cables from PCIE connections on my modular PSU) to play it safe, but it's confused me a little.
 
Last edited:
It's not an overclocked card; in fact GPUz shows its clock is running at 702MHz, which is less than the 738MHz it's specified to run at :confused:

Oh well, it works so I'm not grumbling.
 
Hi Guys,

Wu's completed today - 356960 :D
We gained another 10K on OCForums who are now only 122012 ahead.

Cheers, Simon.
 
We've gained a little more than we should have as some of my results were returned earlier than normal (work PCs are normally set to return work only between 1AM and 7AM).
 
Just been looking through my tasks, I seem to be getting some coming up as error while computing every couple of hours. Any idea why this would be, am I clocking the cards to much or is this normal? Is there any way to find out which card / core is causing the error? :confused:

1377894835 511446143 2 Oct 2009 4:38:07 UTC 2 Oct 2009 5:49:05 UTC Error while computing 1.34 0.01 ---
1377887587 511442711 2 Oct 2009 4:27:24 UTC 2 Oct 2009 5:49:05 UTC Error while computing 1.92 0.01 ---
1377808624 506200690 2 Oct 2009 2:44:01 UTC 2 Oct 2009 4:21:53 UTC Error while computing 1.37 0.01 ---
1377807272 510540704 2 Oct 2009 2:42:39 UTC 2 Oct 2009 4:21:53 UTC Error while computing 1.47 0.01 ---
1377806969 506046171 2 Oct 2009 2:42:18 UTC 2 Oct 2009 4:21:53 UTC Error while computing 1.31 0.01 ---
1377721859 511367944 2 Oct 2009 0:46:36 UTC 2 Oct 2009 2:42:39 UTC Error while computing 1.42 0.01 ---
1377720384 511328311 2 Oct 2009 0:44:26 UTC 2 Oct 2009 2:56:20 UTC Error while computing 97.73 26.60
 
loudbob said:
Well, I seem to be holding fifth spot in the Gauntlet Team.

Can't tell if I'm catching fourth tho......
Click to expand...

well i have just realised that i had forgotten to update the cuda dll's when i updated the driver on my two cuda machines, so now that is corrected i should get a nice boost....

will make up for one of my cuda cards (9600gso) going belly up. not sure what happened to it really, it was added to a dell machine, which started showing a flashing orange light when not on and the power button stopped doing anything. when i put the old non cuda card (7950gx2) back in, it worked again, but when i tried installing the 9600gso card againit started flashing orange straight away, no idea how it works out the card is borked and stops the power button doing anything, but don't want to mess with my other machine right now to test the 9600gso.
 
Who do I talk to for help, I fear addiction to Seti is returning :eek:

i7 @ 3.7 brought on line today inc 275 Gfx card :D

And to celebrate my clients tell me I've cracked 500K on Seti now. Just need everything else to catch up.
 
dekez said:
Just been looking through my tasks, I seem to be getting some coming up as error while computing every couple of hours.
Click to expand...

Looks like the optimized client killing the VLAR WU's

From: 1377894835

<core_client_version>6.6.38</core_client_version>
<![CDATA[
<message>
- exit code -6 (0xfffffffa)
</message>
<stderr_txt>

VLAR WU (AR: 0.010559 )detected... autokill initialised
SETI@home error -6 Bad workunit header

File: ..\worker.cpp
Line: 144


</stderr_txt>
]]>
 
picnic said:
Who do I talk to for help, I fear addiction to Seti is returning :eek:

i7 @ 3.7 brought on line today inc 275 Gfx card :D

And to celebrate my clients tell me I've cracked 500K on Seti now. Just need everything else to catch up.
Click to expand...

I think I have the same problem....having recently got 3 x 260's......and now just waiting for the 300's so I can put together a new tri-sli (or quad + Physx)......all wrapped up in the corsair obsidian case ;)
 
Biffa said:
Looks like the optimized client killing the VLAR WU's
Click to expand...

Ah ok, I see where you got that now. Just googled, seems the optimized client is killing them on purpose because it's a slow running WU on GPUs.

Thanks
 
Erm....

How much of an improvement do the optimised 2.3dll's provide? (especially with 3 x GTX260's)

Cause I just realised I forgot to put them in....D'oh
 
You must log in or register to reply here.
Back
Top Bottom