PPD variance

I'm folding on a few i5 laptops. Typically these get around 5k PPD and have been doing so for months. In the last few days they've dropped to 2500-3500 PPD. I've updated HFM's project list. Right now two I'm watching have P6977 and P6997's.

I used to CPU-Z to verify the CPU's are going at full speed. I've reset the VM's and reboot the host OS' also. Forceasm is on and I don't see any indication assembly options are turning off in the logs.

Is this normal? I don't think my PPD has dropped by as much as HFM indicates.

p6977 is one of the better producing WUs.

Name: WS-18 VM ([email protected])
Path: \\Ws18-ubuntu\ws18-fah\
Number of Frames Observed: 100

Min. Time / Frame : 00:05:09 - 9,377 PPD
Avg. Time / Frame : 00:05:16 - 9,067 PPD

p6997 isn't bad either

Name: Wilson VM ([email protected])
Path: \\Ws17-910\fah\
Number of Frames Observed: 200

Min. Time / Frame : 00:05:24 - 8,733 PPD
Avg. Time / Frame : 00:05:29 - 8,535 PPD

You can always check on the SMP point calculator.

Keep in mind the way HFM calculates ppd. It projects the completion date and time based on the current or RF3 frame times and subtracts the download date/time. If you've turned your machine off for any amount of time, ppd goes way down. One day Harlam is going to add a theoretical or raw ppd column that calculates ppd based on the current frame or RF3 only, disregarding download and offline time, so you can track the affect of hardware and setup changes without having to use the ppd calculator.