FaD and Intel P4's

mrwizer

Senior member
Nov 7, 2004
671
0
0
Has anyone noticed any major differences between Northwoods and Prescotts on FaD? I have three 2.8GHz, only one being a Prescott. Not only is the rating lower on the Prescott (average of about 136 compared to 175), but it does not produce as much by a pretty good margin. I would have thought it would be the opposite, considering that the Prescott has more cache (1MB vs 512KB) and a faster FSB (800MHz vs 533MHz). I guess the architectural differences is the factor at work. Anyone else notice something similar?
 

Wolfsraider

Diamond Member
Jan 27, 2002
8,305
0
76
while I see that I also see a different pattern

For example

My 3.2 does 2, 3 sometimes 4 wu in the same time

or

2x103=206
103 x3=309.9
103x4=412

vs

185 x1 on my 2.8b

now whether this influnces anything I haven't a clue
 

mondobyte

Senior member
Jun 28, 2004
918
0
71
Hi,

There are some very interesting anomalous FaD CPU Ratings phenomena with regards to specific Intel P4 processors and/or some distributions/versions of Linux. The combination of the two provides some extremely "interesting" CPU Ratings [edit] particularly when compared with the actual work done and the time frame of that work [/edit]. I, et al, have documented some extraordinary observations of the behaviour of FaD within this context as well as others. This situation persists even after the CPU Rating "adjustments" by FaD in recent releases and continues to be exploited by some participants.

That is all that I am willing to say on this subject. Do not ask for more information.

[edit]Please note that I do not have any computers, at present, that run Linux or Windows on top of Linux and FaD.[/edit]

mondo
 

mrwizer

Senior member
Nov 7, 2004
671
0
0
I am more curious about the architectural differences in the processors that may be the cause, rather than which OS is being run.

All this talk about CPU ratings make me think about other projects. How do they score progress? I remember Seti@Home (its been a while since I was on the project) was WU's and computer time. CPDN was based off of milestones that had to be reached at set points in the work unit. Are there any projects that work similar to FaD in that they base score off of a synthetic CPU rating related to time?