search

harlam357 / hfm-net

Client Monitoring Application for Folding@Home
34 stars 6 forks source link

Slot PPD and Totals #331

Closed SparklyIsHere closed 4 years ago

SparklyIsHere commented 4 years ago

I find it rather unlikely that several slots have the exact same PPD, to the decimal, so even thou it doesn’t update this more than every 1% the PPD should generally be slightly different.

Rock-001 Slot 1,2 and 5 + Rock-002 Slot 2 are exactly the same 328 983,9

HFMPPD

The Totals also become somewhat different compared to the FAH Control, where you can also see that the Estimated PPD of Rock-001 Slot 1 is 324 417, whereas the HFM software shows it as 328 983,9

HFMTotal

uyaem commented 4 years ago

I can at least explain the difference in numbers between HFM and FAHControl...

Unless you changed the default settings, HFM will average over the last three frames. You can switch this to other averages, or even just the last frame via Edit --> Preferences, tab Options, drop-down Calculate PPD based on in top right-hand corner. FAHControl has a mind of its own and recalculates constantly, even mid-frame. Not sure about the logic there.

Now, as to the points being identical in HFM itself - they are based on the frame time, and if all slots are crunching on the same project (ergo: same base points), and their current (possibly average) time-per-frame is identical (rounded to seconds), it will show the exact same PPD.

SparklyIsHere commented 4 years ago

Okies, thanks, changed the default to ”All Frames” and the same project now shows the variations I would expect to see, as can be seen on the Rock-001 that is currently running 5 instances of the same project 16435.

AllFrames

harlam357 commented 4 years ago

I concur with this statement: FAHControl has a mind of its own and recalculates constantly, even mid-frame.

Glad you got it figured out. :-) Yes, HFM uses an observed approach vs. the predictive approach of FAHControl.