[pb_lee]

(also posted in MacNN's Mac OS X forums)

Hi,

I have been reading a thread on apple's website where people have claimed that their powerbooks are only running at reduced processor speed after the 10.2.8 (v.2) update.
x-bench and other similar programs seem to be reporting processors at their reduced speed.

Here is the link - http://discussions.info.apple.com/W...f2N.2@.599a42a6

Running the following in the terminal seems to verify this (?)
% sysctl hw.cpufrequency
hw.cpufrequency: 533333332

just wondering what other people think... is your computer running slower after the update, or is it actually the sysctl command that is screwy and reporting false results (therefore meaning x-bench will display the incorrect result!?) and yes i have tried this with both the "highest" and "reduced" settings in the energy saver control panel.

lee

[Pierre B.]

Originally posted by pb_lee:

just wondering what other people think... is your computer running slower after the update, or is it actually the sysctl command that is screwy and reporting false results (therefore meaning x-bench will display the incorrect result!?)

I think it is a bug in reporting the correct CPU frequency. You can run tests to verify this and see that the computer is faster when set to highest, despite the report of xbench or sysctl (even the CPU preference pane-you have to install CHUD-reports incorrectly). It looks like a bug at kernel level.

For example, you can download the code from this Appleinsider thread and test it under the reduced and highest settings for the processor speed. The settings make no difference in sysctl or xbench report, but in actual code runs they do make.

[pb_lee]

thanks for the reply pierre b!

its good to know that nothings actually wrong!

lee

[amazing]

There's a new version 1.1.3 of X-Bench out, here's what the author says:

"I believe this is resolved in 1.1.3. For the curious, this was caused by the speed stepping feature of the processor, which engages more aggressively on 10.2.8 for whatever reason. In other words, everything was normal with the machines in question -- Xbench was just grabbing the *current* clock frequency, rather than the maximum possible on the processor."

http://spiny.com/forum/viewtopic.php?t=86

PS: The aggressive processor shifting in 10.2.8 rev b would also account for the improved battery times...

also, if your PB seems to run slow at the reduced processor rate, then re-applying the 97 MB combo 10.2.8 update is said to fix that. Just don't forget to always repair permissions after any update...