The WALLCLOCK time-space between checkpoints is dependent on the speed of the CPU and the idle time the science gets allocated on the system.

As for not sticking to the throttle for sure the process will not adhere to the throttle until the model setup is done. It read like this takes an extra-ordinary long time.

My HCMD crunchers continue to behave normal albeit longer, but with the project having a lower than low priority setting, really interested also to hear if 100% CPU time truly does clock up time in Taskmanager i.e. if it runs 20 hours in the main screen, how much time has it accumulated in the Taskmanager for wcgridHCMD.exe process. A Possible AV interference after an update is always lurking around the corner.

Anyone reporting in, plz include:

Windows OS:
Time Wall-clock per Agent Main Screen:
Accumulated CPU time (per Taskmanager, if run uninterrupted):
Percent progress per Agent Main Screen:
Percent progress per Agent Graphics Screen:
Throttle % setting per the Agent Graphic Screen:
CPU % indicated per Taskmanager:
AntiVirus Version and update level:
Any other observation:

thanks

We are looking into the cpu throttle and snooze issues with some of the longer running HCMD wus. Stay tuned for updates.

Thanks,
armstrdj

Demon999 et al,

WCG will not stop long work units just because they are long, nor is the number of work units put thru overall influencing the decision to pull a batch of work. The present set - HCMD JOBS ONLY - are simply tough. They run on the WCG reference machine 2ghz CPU with 100% throttle for 10 CPU hours. On my C2D they do about 13-14 CPU hours. Throttled down on a slower machines they will get longer. That said, the project scientists need those jobs done, so please hang in there!

The higher weight put on the WCG(UD) agent side of the project for HCMD is compensated by the BOINC agent side. Overall we still did 158,000 work units on Thursday. The global work unit count can be followed on the statistics page: http://www.worldcommunitygrid.org/stat/viewGlobal.do

Thank you for your patience and perseverance to help conclude this project successfully.

sekerob

Tx Sekerob,

I thought this is a defect in the WCG as I only received long running workunits on WCG and not on BOINC (still the case).

But your post is clarifying it and I am happy to run these long ones through. I also noticed that these long ones always have a similar graphics which lead me to believe that the scientists are using these for confirmation on something with a bigger scale.

Wow, I didn't realize the jobs got that big. I just checked my oldest machine (1.2GHz) and it's been working on a HCMD unit for 13 hours and only 19%... As long as it keeps crunching away I'll be happy!