Still a problem with the “enforce_delay_bound� option. Message from server: No work sent (won't finish in time)

Ok, let’s analyze this.

So the sever dose this estimate:

A WU will take 28H to crunch.
BOINC can use my CPU 11% of the day.

So to return a WU in time i will need at least a 10-day deadline. (9.69) So i will have no chance in hell to crunch that WU in less than 7 days.

Now in reality things are worse, a WU will need 40H not 28H to crunch. So if BOINC only can use my CPU 11% of the day, the deadline would have to be at least 15 days (15.15) so i could return a WU in time. Can we agree on this so far?

Now my computer must have good connections with some higher power (it should have because it have already been frown in the Dumpster once) or the CPU must be located in a hole in the space-time continuum, because it haven’t returned a WU late yet and has an Average turnaround time of 5.3 days.

But before we ask for a grant to explore this and make plans for a BOINC project aimed at exploiting this, we might examine if there could be an alternative explanation for this.

The estimate for the crunch time of the WU is wrong, but it is to my advantage so it isn’t the problem here. That only leaves the estimate of how much CPU-time BOINC will get in a day. If i understand it correctly the estimate of the available CPU-time (running_frac) is calculated by multiplying “the percentage of the day computer is on� with “the percentage of CPU-time available to BOINC then it’s on�.

In my case it is:
Computer on 34.6% of time
BOINC on 34.6% of that

Now this looks suspicious to me. Both values are 34.6%. Since we know something is wrong, my guess would be that one of these values is a copy of the other. That the Computer is on 34.6% of time seams fairly accurate. My guess would be something lower then 50%. That BOINC would only have access to the CPU 34.6% of time seams very low to me. This computer is manly used to surf the web, reading BOINC message boards and to listen to internet-radio through winamp (18% CPU use). So i would expect a value around 80%.

But this doesn’t explain it all. If the computer were on 34.6% BOINC would on average have access to the CPU 97-98% of the time to be able to have a turnaround time in 5.3 days. This isn’t possible on my machine, so let’s play around with my estimates. If my machine was on 50% of the time and BOINC had access to the CPU 80% then running_frac would be 40%. That is fairly close to 34.6%. Now lets say my computer was on 43.25% and BOINC had access to the CPU 80% then the running_frac would be 34.6%.

So let’s run the numbers:
A WU will take 40H to crunch.
BOINC can use my CPU 34.6% of the day.

40/(24*0.346)=4.82

Since my last WU was done in 4 days 10H 10 min and my average turnaround time is 5.3 days, this isn’t such a bad estimate.

So my guess is that the values for “ Computer on� and “ BOINC on� is actually running_frac. So then the server calculates running_frac it actually gets squared.

Now how can we test this theory? Well that’s easy. Lets look for computers that have the same value for both “ Computer on� and “ BOINC on�. Fist of, does my computer have that? If not that’s interesting because that was what the server reported. If it has, is it all computers or just, lets say all windows 98 users?

Bruce, do you think i can be on to something hear?
Thanks for the fast replay as always.

-----------------

The reason my machine have 0.15 seconds of remaining work for E@H is because i have set the “contact server every� value to 0.02. Then running application version 4.75 the client downloaded a new WU 5 min before the old finished (value 0.01). With application version 4.79 it seams that the calculation of remaining time is not so accurate for the last minutes. I think i will try 0.05 the next time.

It seams that i am not the only one noticing that both “ Computer on� and “ BOINC on� hase the same value. Debugas spotted it also. Fj has at least one computer with the values being the same.

http://einsteinathome.org/node/188003

Lets se if we can sort this out.

In addition to the problems i got, you got the following:

1. A WU is estimated to take 41H but in reality only takes 24-28H
This is a problem if it is correct. My computer say it will take 40H but the server say it will take 28H. It would be interesting to se what time the server estimates that your WU will take.

2. Since you run multiple projects only half the CPU time is available to each project.
The server isn’t miss adjusting, it does what it is supposed to do in this case by halving the calculated time, because only half the time is available to this particular project.

edit

Now if i only could get the dyslexia under control i wouldn’t have to delete things so often.
Well i could always blame it on that it is getting late hear.