My rig is a Windows 10 with NVidia 1080 with the latest driver 461.40 but every WUs ends with a message like those ones

2/10/2021 5:32:17 PM | MLC@Home | Output file ParityModified-1611721867-28196-2_0_r1807205113_0 for task ParityModified-1611721867-28196-2_0 absent
2/10/2021 5:32:17 PM | MLC@Home | Output file ParityModified-1611721867-28196-2_0_r1807205113_1 for task ParityModified-1611721867-28196-2_0 absent

Just the same problem with a Linux computer (NVidia 970)


Is there a project admin who can say something about this problem ????

I had problems with Win10 and crunching the gpu units on my laptops Nvidia 1660 but fixed all of them by stopping Boinc, downloading the latest all-in-one Visual-C-runtimes file and extracting and running all of them and then after restarting Boinc resetting MLCathome and now I am crunching gpu files just fine so far. I got the Visual-C-runtime files here:
https://www.techpowerup.com/download/visual-c-redistributable-runtime-package-all-in-one/

Note you shouldn't need to download the runtimes, they should be included in the windows download. But windows it a neverending source of random configs that work for most people but fail for some.

In this case, your computer seems to be erroring out with

<core_client_version>7.2.42</core_client_version>
<![CDATA[
<message>
app_version download error: couldn't get input files:
<file_xfer_error>
  <file_name>libcudnn_adv_train.so.8-pt17cuda102000</file_name>
  <error_code>-200 (wrong size)</error_code>
</file_xfer_error>

</message>
]]>

So it looks like there was a problem downloading your client. I realize we're a month late replying, but if you could reset the MLC@Home project, which will re-download the client, that might help clear the error.

Similar Problem on my side with nvidia gtx 960. Die client errored out with:

<core_client_version>7.16.11</core_client_version>
<![CDATA[
<message>
(unknown error) - exit code 3221225785 (0xc0000139)</message>
]]>

-1073741511 (0xC0000139) STATUS_ENTRYPOINT_NOT_FOUND

No success with update to latest nvidia drivers, but after reset the project and re-download all runtimes from server with the boinc-client, the application now running.

Regards Odi

---

I switched back to running MLC after a while away. I wasn't getting any of the error messages mentioned here, but I was getting run time estimates of 10+ DAYS on my GTX 1060, so I reset the project and now it seems to be running properly.

I switched back to running MLC after a while away. I wasn't getting any of the error messages mentioned here, but I was getting run time estimates of 10+ DAYS on my GTX 1060, so I reset the project and now it seems to be running properly.

I spoke too soon. It did error out after running for nearly 10 hours. I guess I'll go back to running something else on the GPU for now...

https://www.mlcathome.org/mlcathome/result.php?resultid=4604946

Is this duration normal for the GPU App ...

4866700 2588430 11 Apr 2021, 8:09:58 UTC 11 Apr 2021, 14:45:37 UTC Terminé et validé 23,585.31 21,438.87 4,160.00 Machine Learning Dataset Generator (GPU) v9.75 (cuda10200)
windows_x86_64