9070XT and Bad WU

[u7cappy]

I appear to be running in to a LOT of failed work units, failing roughly 32%. Digging through logs shows a similar message typically,

Code: Select all

09:47:44:I1:WU44:An exception occurred at step 150343: Particle coordinate is NaN.  For more information, see https://github.com/openmm/openmm/wiki/Frequently-Asked-Questions#nan
09:47:44:I1:WU44:Max number of attempts to resume from last checkpoint (2) reached. Aborting.
09:47:44:I1:WU44:ERROR:114: Max number of attempts to resume from last checkpoint reached.
09:47:44:I1:WU44:Saving result file ..\logfile_01.txt
09:47:44:I1:WU44:Saving result file science.log
09:47:44:I1:WU44:Saving result file state.xml.bz2
09:47:44:I1:WU44:Saving result file xtcAtoms.csv.bz2
09:47:44:I1:WU44:Folding@home Core Shutdown: BAD_WORK_UNIT
09:47:44:E :WU44:Core returned BAD_WORK_UNIT (114)

I'm running two AMD gpu's. One 6700xt in a separate rig, which has yet to fail a unit. Then my 9070xt.. which at 32% failing rate feels like its just wasting power.
Card has never had any stability problems, running stock with an aggressive fan curve. Does not exceed 82C hotspot temp.

Is there anything I can do to improve my odds here?

[muziqaz]

Those projects will be disabled for 9070xt.
In testing my 9070 on Linux did not see any issues, except with one WU. Probably something in Windows driver is triggering something with those projects
Thanks for the report

[u7cappy]

That was fast, thank you!