Consistantly Lowest

mlcudd
mlcudd
Joined: 22 Jan 05
Posts: 26
Credit: 51,808
RAC: 0
Topic 188476

Hi All,
I have gone over the other posts and Paul Bucks documentation, but have been unable to determine why I seem to have the Lowest 'claimed Credit" on all the WU's that I do. It seems to happen the most on my Athlon box, but I even noticed it on my Celeron Box today. Anyone have any ideas, or is that just the luck of the draw?

Respectfully,

Rocky Cudd


WinXP SP2 Boinc - 4.45
Have A Great Day!

ben
ben
Joined: 9 Feb 05
Posts: 36
Credit: 1,663
RAC: 0

Consistantly Lowest

Can't say as I can't see your computers credit listings.

mlcudd
mlcudd
Joined: 22 Jan 05
Posts: 26
Credit: 51,808
RAC: 0

sorry about that, already

sorry about that, already changed! Thanks

Regards,

Rocky


WinXP SP2 Boinc - 4.45
Have A Great Day!

ben
ben
Joined: 9 Feb 05
Posts: 36
Credit: 1,663
RAC: 0

Looks to me, from short

Looks to me, from short survey, that you seem to be being granted higher credits than you are claiming...which we all agree is a good place to be.
(for the first page of all your 3 hosts)

It all depends on the other hosts benchmark scores...which from so so many discussion threads we know people have objections to. However you seem to be coming out ahead.

Some others claim much lower ammounts per second as they are using Linux...but again the einstein client on linux seems to crunch much slower also.

#1 host
Claim Grant
73.43 100.26 -more
73.43 97.02 -more
76.48 90.40 -more
77.68 97.94 -more
76.38 82.47 -more
79.16 97.93 -more
80.84 99.90 -more
78.32 92.26 -more
81.70 85.61 -more
78.66 92.27 -more
78.55 92.68 -more
79.96 79.96
81.68 88.52 -more

#2 host
67.17 102.05 -more
66.87 pending
66.98 95.40 -more
67.12 67.12
67.10 70.81 -more
67.29 97.87 -more
67.42 98.83 -more
69.83 67.81
67.88 100.79 -more
66.97 77.93 -more
70.45 70.66
63.94 94.12 -more
70.15 68.80

#3 host
86.78 104.04 -more
86.20 86.91 -more
61.02 pending

mlcudd
mlcudd
Joined: 22 Jan 05
Posts: 26
Credit: 51,808
RAC: 0

Hi Ben, I do agree with you

Hi Ben,
I do agree with you that I am recieveing more than claimed credit,and I am happy about that. That was what I was wondering about. So basically what you are saying is that different operating systems crunch the same WU's at a different pace, which in turn directly affects the credit.

Thank You,

Regards,

Rocky


WinXP SP2 Boinc - 4.45
Have A Great Day!

ErichZann
ErichZann
Joined: 11 Feb 05
Posts: 120
Credit: 81,582
RAC: 0

are you using boinc 4.25 ? if

Message 8200 in response to message 8199

are you using boinc 4.25 ? if you are the others are still maybe using version 4.19 and have higher claimed credits

Paul D. Buck
Paul D. Buck
Joined: 17 Jan 05
Posts: 754
Credit: 5,385,205
RAC: 0

Rocky, > I do agree with

Message 8201 in response to message 8199

Rocky,

> I do agree with you that I am recieveing more than claimed credit,and I am
> happy about that. That was what I was wondering about. So basically what you
> are saying is that different operating systems crunch the same WU's at a
> different pace, which in turn directly affects the credit.

The bit about credit is a weird "dance" with several factors that intertwine in nice odd ways to make generalizations hard. However, I am going to give it a shot ...

True calculation speed is a factor of:
1) CPU type
2) clock speed
3) compiler optimization/code efficiency

System efficiency is a factor of:
1) system load
2) Operating System
3) True calculation speed

On top of that we throw...

Benchmark score affected by true calculation speed.

So, since credit is based on CPU time and benchmarking score any change or difference in any of the above factors affects speed.

Ignoring the "luck of the draw", which you may also have been affected by ...

You may have had lower than expected benchmark scores and so this lowers the claim. There has been a historical issue with the benchmark not being consistent across the OS and we get to stir the pot.

There is discussion in other threads about running the Windows binaries on Linux using the Wine program and some people are getting better performance and/or better benchmark scores. This is an indicator of the Operating System and compiler affecting calculation speeds.

In the older versions of BOINC the programs were compiled using the Microsoft compiler for windows and, i think, the gcc compiler for Linux and the Apple gcc for the Macintosh.

mlcudd
mlcudd
Joined: 22 Jan 05
Posts: 26
Credit: 51,808
RAC: 0

Thank You Metalwarrior and

Thank You Metalwarrior and Paul,
First off I am using 4.25. And my benchmarks changed when I went to 4.21, however they have not really changed since going through the other development clients until I got to 4.25, where they went down a fair amount. I did not take this into consideration. Paul, I thought about what you just wrote, and if I am understanding it correctly, the machine I used the most for all my daily activities would process a WU differently than the other 2 machines that are used basically just for crunching. Where I got confused is, my slowest machine, the one that runs Windows ME crunches a Einstein WU in about 27 hours average. The "estimated completion time is 14 hours. I figured this was way off because of the low Mhz Celeron processor, and with the increased processing time the claimed credit is always lower than the Granted credit.
I apologize if this sounds redundant. I am not compaining, I am just trying to get a better grasp on the project.
Another question comes to mind, Would my slower processor machine be better to crunch a long term Climate WU, or a shorter project WU? Would it be better if I had only one project attached to the slowest box?
Again if the questions have been addressed before, I apologize.

Respectfully,

Rocky


WinXP SP2 Boinc - 4.45
Have A Great Day!

Paul D. Buck
Paul D. Buck
Joined: 17 Jan 05
Posts: 754
Credit: 5,385,205
RAC: 0

> I apologize if this

Message 8203 in response to message 8202

> I apologize if this sounds redundant. I am not compaining, I am just trying
> to get a better grasp on the project.

As long as it is not duplicating redundent we should be Ok. :)

All I was trying to convey is that it is not as simple as it looks. Everything touches everything else. It is like my question about should I go with a dual Operton/Athlon or with a dual Xeon. The Xeon gets beat on run time but seems to win on throughput ...

> Another question comes to mind, Would my slower processor machine be better
> to crunch a long term Climate WU, or a shorter project WU? Would it be better
> if I had only one project attached to the slowest box?

I would say yes, I would run that for SETI@Home and Predictor@Home. Mostly from a paranioa standpoint. If there is a problem with the machine you can lose more work. From a BOINC System standpoint it matters not at all. But I would not run Einstien@Home or CPDN from the standpoint of EAH has the short deadline and CPDN, well, you still may fail to make the deadline.

I would still run 2-3 projects on it just so you don't run out of work ... though even a week might not be enough project downtime for that machine to notice ... :)

> Again if the questions have been addressed before, I apologize.

John Wayne said once, "Don't apologise, it is a sign of weakness" ... :)

mlcudd
mlcudd
Joined: 22 Jan 05
Posts: 26
Credit: 51,808
RAC: 0

Thanks Paul, That clears it

Thanks Paul,
That clears it up for me. I appreciate the quick response. Now at least I got the concept and have to switch a couple of programs but no big deal.

Regards,

Rocky


WinXP SP2 Boinc - 4.45
Have A Great Day!

lysdexia
lysdexia
Joined: 9 Mar 05
Posts: 97
Credit: 17,013
RAC: 0

consistently redundant

consistently
redundant


"My other computer is a virus farm."

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.