just wonderd that there are host with an RAC=Recent average credit von over 300 with an P4 3.0...

my P4 3.2C running at 16 x 220 = 3520 at 2.5 3 3 8 on an I875 Mobo has only an RAC from 134.99 :confused:

and it is running 24/7 Hours a Day...
Greetings from Germany NRW
Ulli [/url]

the mystery of RAC
will not be solved

If you merge identical host to a "larger" you get a
"superkomputer", this will not have an influence in dayly rate.

sharing confusion?

2 identical p3.0 800 /512, maybe a difference less 2%


p3000 171.09 3577.29 Intel(R) Pentium(R) 4 CPU 3.00GHz Pentium
p3002 150.88 2537.17 Intel(R) Pentium(R) 4 CPU 3.00GHz Pentium

alles verstanden? everything clear?

p2500 187.81 1330.89 AMD Athlon(tm) XP 3200+ Pentium

cheap amd p2500 running at 2291 MHz instead 2000 MHz, 256 MB

just installed boinc, attached 2 project, nothing tweaked

u asked for RAC value, here are some.
But please do not ask for a clear anwser and don't wonder to much
gibt Kopfweh, gives only headake.

I'm sure you have forgotten to add the cpus temperature to your actual RAC value.

as more you try to solve this mystery as deeper the confusion will grow

ric

The RAC can move way up when other users return the third crunch of WU's that you had crunched the 1st or 2nd of. Due to that it can "seem" almost random
<A>

RAC means you have to deliver the WU,s at a frequent rate.
So if like me you do say 15 WU's a day then it takes you a couple of days to get connected to SETI and to get more your RAC number will drop.

today the 21/07/04 my credit reads.
Total credit 7194.36
Recent average credit 372.60
No WU's for 3 days.

on the 19/07/04 it was
Total credit 6699.37
Recent average credit 418.63
Daily connections

on the 16/07/04 it was
Total Credit 5493.30
Recent average credit 396.51
Daily connections

So there you go..... clear as mud.

Allow me to add to the confusion. These are some examples of mine:

2.4Ghz P4 256MB RAM - RAC: 102.84 Total Credit: 2392.44 WinXP
1.7Ghz P4 256MB RAM - RAC: 100.69 Total Credit: 2277.69 WinXP
850Mhz P3 128MB RAM - RAC: 73.41 Total Credit: 1539.14 Win2k

and finally....

3.2Ghz HT P4 1GB RAM - RAC: 72.11 Total Credit: 917.45 WinXP

All have been running the same length of time (within 2 days). Of course, I could have simply said I have no idea how RAC is calculated.....


<a>

the short list shows the values are more like random

RAC should view something like a daily rata of processed WU.

But this can't be actual.

the measurement is asyncron, your contribution syncron.


In addition there must be a large backlog in UCB,
stagnation in total credits and decreasing RACs

a backlog full of "claimed credits".

guess, there is no chance of having an "online" RAC returned

>I have no idea how RAC is calculated.....
it's not clear how a cobblestone is calculated, found on the net 3 "formulas"
every gives an other result (large smile)

if this could cleared one day, then a tool can calculate the estimated
possible credits per 24h due the CPUs possibilities.

ric

Yep, clear as mud

RAC has never been easily explained. Even the definition I have for it is about as clear as mud.

RAC is also complicated if you are running two projects (as I am) because when you swtich from processing on one project to another this causes the RAC to start to decay one that project while the new project's RAC should be increasing.

I was hoping that the secondary sites that do the stats would do a "merge" of the cross project values so you would have a better "feel" for the daily numbers.

A long time ago in the Beta I raised some questions on the cross project science application(s) and the interactions. Though I will admit that my focus there was on the BOINC Software on the participant's computer and fortunately the problems there seem to be minor (though I have had crashes on predictor's science application). But I also asked the questions about the server side processing and the linking of the back-end's processing across projects.