I couldn't find another thread for this problem. Mods feel free to delete this if it's already in another thread and please point me to it. Thanks.

Just attached to this project early this morning and every work unit has had an error. I've set BOINC to no new tasks until this problem is worked out. Can someone please give a newbie some advice? Thanks

stderr out

<core_client_version>6.6.38</core_client_version>
<![CDATA[
<message>
Incorrect function. (0x1) - exit code 1 (0x1)
</message>
<stderr_txt>
# Using CUDA device 2
# There are 3 devices supporting CUDA
# Device 0: "GeForce GTS 250"
# Clock rate: 1.84 GHz
# Total amount of global memory: 1073741824 bytes
# Number of multiprocessors: 16
# Number of cores: 128
# Device 1: "GeForce 9800 GT"
# Clock rate: 1.62 GHz
# Total amount of global memory: 536870912 bytes
# Number of multiprocessors: 14
# Number of cores: 112
# Device 2: "GeForce 9800 GT"
# Clock rate: 1.62 GHz
# Total amount of global memory: 536870912 bytes
# Number of multiprocessors: 14
# Number of cores: 112
MDIO ERROR: cannot open file "restart.coor"

</stderr_txt>
]]>

By the way, I have about 1000 SETI work units in cache. If the fix for the problem might trash my cache, It'll be several days before I can try it. I don't want to abort all those units.

Thanks again.

So far I've found a couple possible fixes for my problem. I'll let my cache run dry, uninstall and detach projects and use driver-sweeper. Then install the newest version of nvidia graphics driver, newest version of BOINC and give it another try.

This will take a few days. Meanwhile, any additional advice will be appreciated.

Thanks

Well, I finally drained my SETI cache and reinstalled everything. Then used driver sweeper as planned. So far my GTS 250 is crunching, good deal..

BOINC didn't see my other two 9800 GT's so I made a cc_config.xml file and dropped it in the data folder. That worked and they started crunching but the units failed almost immediately with the same error as before. At least my GTS 250 works now... Still a big disappointment to me.

So far the only project that utilizes all the power of my computer is SETI.

I have all the latest drivers for my nvidia motherboard, gpu's and even the latest version of bios.

The config file consisted of this:
<cc_config>
<options>
<use_all_gpus>1</use_all_gpus>
</options>
</cc_config>

And I also had to use dummy plugs like I had to for SETI to recognize all the GPU's.

Anyone have any advice? I'm willing to try anything at this point.

One of my units that had an error claimed 4,531.91 credits for 8.21 seconds of crunching! Several others claimed ridiculous amounts of credit too.

http://www.gpugrid.net/result.php?resultid=1450434

Well, I finally drained my SETI cache and reinstalled everything. Then used driver sweeper as planned. So far my GTS 250 is crunching, good deal..

BOINC didn't see my other two 9800 GT's so I made a cc_config.xml file and dropped it in the data folder. That worked and they started crunching but the units failed almost immediately with the same error as before. At least my GTS 250 works now... Still a big disappointment to me.

So far the only project that utilizes all the power of my computer is SETI.

I have all the latest drivers for my nvidia motherboard, gpu's and even the latest version of bios.

The config file consisted of this:
<cc_config>
<options>
<use_all_gpus>1</use_all_gpus>
</options>
</cc_config>

And I also had to use dummy plugs like I had to for SETI to recognize all the GPU's.

Anyone have any advice? I'm willing to try anything at this point.

Well you have the correct cc_config for using all GPU's. You also have up to date drivers which is usually a common problem. The message you are getting usually suggests a cuda issue.

All the wu that got an error seem to have been run on the 9800GT's. There was one there that started on the GTS250 and then when its restarted its on a 9800GT, but then errored out.

You aren't using an app_info are you? Are there any messages in the BOINC log? Have you tried a detach/reattach project after updating drivers?

I'm running 190.62 drivers under XP. Maybe you could try this version which supports cuda 2.3

Usually your type of error is due to overclocking either by manufacturer or by you. However, it is strange that it happens to all WUs.
Do they crash immediately or run for a while?

gdf

One of my units that had an error claimed 4,531.91 credits for 8.21 seconds of crunching! Several others claimed ridiculous amounts of credit too.

http://www.gpugrid.net/result.php?resultid=1450434

Thats why its claimed credit. The granted figure is zero.

Unfortunately Vista and Win 7 require dummy plugs, something that XP doesn't need. It supposedly will be addressed by a later driver version from nvidia, but don't hold your breath.

At this point don't buy GTX295, a new G300 (coming out in a couple of months) should be faster than both cores of a 295.

gdf

oh, one more thing... the two errors for Nov 3rd are from me trying one last time to get my bad GPU working. I put an SLI Bridge (Think that's what it's called) between the good and bad GPU to see if it would make it work. It's officially dead. Won't even run the monitor if it's the only one in the computer. But like I said, the bad GPU is totally my fault and has nothing to do with the project. I don't want anyone thinking GPU grid will kill their video cards...

A GTX295 consumes more than a G300. It is unlikely to be so expensive with the competition of ATI.

A GTX295 consumes more than a G300. It is unlikely to be so expensive with the competition of ATI.

gdf