You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I noticed a strange behavior in OhmGraphite: Values were starting out mostly correct (sometimes even wrong after reboots) but over time were increasingly lagging behind in comparison to other tools like LibreHardwareMonitor, FanControl or HWInfo64. The latter tools were all reporting the same values at the same time. We eventually figured out that setting OhmGraphite's reporting interval to 1 second instead of 5 seconds solves the problem.
Now, the problem also occurs in the LibreHardwareMonitor application, if the interval is something different than one second. I could even "catch-up" to the values reported by other tools, if I changed the interval from 5 seconds down to 250ms:
I set it to 5 seconds in the LibreHardwareMonitor UI, then increased and decreased the actual temperature value (as reported by FanControl) from 29 up to 32, and back to 29. No changes in the UI. You can, however, set the refresh interval to 250ms which works like a "fast-forward" in this case and I was able to observe the increase and decrease after the fact in a sort of timelapse!
I wasn't able to find any documentation on the protocol introduced with #938 but it feels like the buffer is growing constantly instead of getting overwritten by the Quadro? Only guesses on my part, though.
Happy to help create more data points for you to find the issue, or test out any ideas you might have.
The text was updated successfully, but these errors were encountered:
I noticed a strange behavior in OhmGraphite: Values were starting out mostly correct (sometimes even wrong after reboots) but over time were increasingly lagging behind in comparison to other tools like LibreHardwareMonitor, FanControl or HWInfo64. The latter tools were all reporting the same values at the same time. We eventually figured out that setting OhmGraphite's reporting interval to 1 second instead of 5 seconds solves the problem.
You can find the thread here: nickbabcock/OhmGraphite#440
Now, the problem also occurs in the LibreHardwareMonitor application, if the interval is something different than one second. I could even "catch-up" to the values reported by other tools, if I changed the interval from 5 seconds down to 250ms:
I wasn't able to find any documentation on the protocol introduced with #938 but it feels like the buffer is growing constantly instead of getting overwritten by the Quadro? Only guesses on my part, though.
Happy to help create more data points for you to find the issue, or test out any ideas you might have.
The text was updated successfully, but these errors were encountered: