Josh Saddler - Latest comments on Hardware success!

In response to: Hardware success!

energyman [Visitor] — Thu, 30 Oct 2008 02:39:49 +0000

oh, not in the kernel config ;)

afaik the problem is that when pat is turned on by the kernel and the driver there is some bad interaction. Pat support in kernel, deactivated at command line and turned on by the driver works well (for me...)

In response to: Hardware success!

Josh Saddler [Member] — Wed, 29 Oct 2008 19:34:51 +0000

@Donnie & Jeremy:
I already use cpufreq to run "ondemand" the whole time. However, that's not the same as getting lower heat and saving a bit of energy regardless of frequency, which can be accomplished by lowering the voltage to the CPU.

On the old motherboard, I dropped the voltage from the default 1.35V (IIRC) all the way to 1.150. Could have gone a bit lower, judging by other success stories, but I decided not to push my luck. As it was, I saw about ~5C decrease in temps regardless of load.

@energyman:

I already knew about that before I even installed the drivers, so I made sure to have PAT disabled in my kernel config ahead of time. Didn't make a difference.

In response to: Hardware success!

energyman [Visitor] — Wed, 29 Oct 2008 18:29:39 +0000

when 3d stutters with newer ati hardware, you need to pass nopat as a kernel boot option.

In response to: Hardware success!

Jeremy Olexa [Visitor] — Wed, 29 Oct 2008 15:08:47 +0000

1) userspace utilities are very easy to use wrt cpu freqs (as Donnie mentioned)
2) Coolness about the new router. Sounds neat.

In response to: Hardware success!

Donnie Berkholz [Visitor] — Wed, 29 Oct 2008 14:35:45 +0000

To save power and heat, why resort to fiddling with the BIOS? You should be able to use the kernel cpufreq on even desktop CPUs nowadays.

In response to: Hardware success!

Donnie Berkholz [Visitor] — Wed, 29 Oct 2008 14:34:28 +0000

I wonder if the guy at D-Link who knew Gentoo quit, and that's why it got EOL'd. =)