New WUs 16% longer in CPU Time but same point value?

qdemn7
qdemn7
Joined: 20 Feb 05
Posts: 12
Credit: 3414228
RAC: 0
Topic 193092

What's going on with these?

I've completed 4 of these with 2 more processing:

34627510

34634461

34634461

34637366

34640067

34643078

As a comparison this in an average 667 point WU:

34620956

Stick
Stick
Joined: 24 Feb 05
Posts: 790
Credit: 33268922
RAC: 7937

New WUs 16% longer in CPU Time but same point value?

The new v4.38 app is slower. Read this thread.

Quote:

What's going on with these?

I've completed 4 of these with 2 more processing:

34627510

34634461

34634461

34637366

34640067

34643078

As a comparison this in an average 667 point WU:

34620956


qdemn7
qdemn7
Joined: 20 Feb 05
Posts: 12
Credit: 3414228
RAC: 0

OK, thanks for the help.

OK, thanks for the help.

archae86
archae86
Joined: 6 Dec 05
Posts: 3161
Credit: 7275111720
RAC: 1885513

This does not answer your

This does not answer your already answered question, but I'm just curious.

My Q6600 running 3.006 GHz is needing about 150.5 CPU seconds/credit. Yours is needing only about 131.

So are you overclocking this 2.4GHz part to about 3.45? While my overclock is conservative, my motherboard won't go your speed, and my part would require pretty high CPU voltage and burn a lot of power to go that fast.

th3
th3
Joined: 24 Aug 06
Posts: 208
Credit: 2208434
RAC: 0

Im positive that qdemn is a

Im positive that qdemn is a serious overclocker :) I remember seeing his E6600 in the top20 computers, half a year ago or longer, no other dualcore above him in periods, must have been running at 3.6 or more at the time.

googloo
googloo
Joined: 11 Feb 05
Posts: 43
Credit: 13411395
RAC: 785

I hope some adjustment in

I hope some adjustment in credit is planned?

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3522
Credit: 762427658
RAC: 1078934

RE: I hope some adjustment

Message 72769 in response to message 72768

Quote:
I hope some adjustment in credit is planned?

There's a fixed credit number for each individual workunit that will be granted by the server side BOINC processes no matter what platform or version of the client is used.

The debug code that slows down the newest app was included to identify one specific bug, when this is found the specific debug code will be removed again and the app will run faster again. Hopefully this will happen rather sooner than later.

CU

H-BE

googloo
googloo
Joined: 11 Feb 05
Posts: 43
Credit: 13411395
RAC: 785

Thanks for the explanation,

Thanks for the explanation, Bikeman. However, I don't think it's fair to increase the time for a run by more than 10% without increasing the credit. Whenever rosetta or SETI get running again, I'm out of here until the apps are fixed or until S5R3. I usually don't base my decisions on credit, but fair is fair.

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3522
Credit: 762427658
RAC: 1078934

RE: Thanks for the

Message 72771 in response to message 72770

Quote:
Thanks for the explanation, Bikeman. However, I don't think it's fair to increase the time for a run by more than 10% without increasing the credit. Whenever rosetta or SETI get running again, I'm out of here until the apps are fixed or until S5R3. I usually don't base my decisions on credit, but fair is fair.

Well, if you are strictly following this line of argumentation, then every speed-up of the app should be reflected in a credits change as well (by lowering them), right? It's unavoidable that during the evolution of the apps, they get faster most of the time and slower sometimes.

CU

Bikeman

rbpeake
rbpeake
Joined: 18 Jan 05
Posts: 266
Credit: 1149552797
RAC: 660957

RE: Well, if you are

Message 72772 in response to message 72771

Quote:

Well, if you are strictly following this line of argumentation, then every speed-up of the app should be reflected in a credits change as well (by lowering them), right? It's unavoidable that during the evolution of the apps, they get faster most of the time and slower sometimes.

CU

Bikeman


Ha, ha, good point! ;) Although perhaps one could argue that points are "sticky upwards", meaning that when you are dealing with cruncher psychology, upwards in points is always OK, but downwards is never really OK at all! ;)

But I do follow your argument, except I think in the case of BOINC there is a "symmetry violation" to your argument! ;D

Comment viewing options

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