![]() |
Re: A faster N810??
Quote:
|
Re: A faster N810??
Quote:
|
Re: A faster N810??
Quote:
|
Re: A faster N810??
Quote:
Basically, just do this as root: Code:
echo performance > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor Fennec is almost usable! |
Re: A faster N810??
Did you test what it does to battery life?
|
Re: A faster N810??
Quote:
Speaking from a Windows Mobile point of view, on-demand and performance are the same thing, on-demand just throttles the CPU down when not in use, performance doesn't. How is it different in Maemo? Thanks for mentioning liqbase, looks cool. I'm afraid Fennec is a disaster, unusable. |
Re: A faster N810??
ioioio:
yes, this just prevents the CPU from throttling down from 397.46 to 164.36 Mips (those numbers are according to my home applet). lcuk and others discovered that turning off the throttle makes apps run faster, probably because the CPU throttle is too aggressive for CPU-intensive apps. As for command line phobia: Put that line into a Personal Menu item and set it to run as root. Problem solved! Matan: I've seen a very mild battery life decrease. So mild that I'm not sure it exists. It isn't very significant, because I don't believe the CPU speed is a big determiner for battery life. If anything, a faster CPU should help "race to idle," no? In any event, you can set your processor back to on-demand, just by replacing the "performance" in the command line with "ondemand". |
Re: A faster N810??
Quote:
The flip side, of course, is that you'll get a delay while the CPU ramps up for more demanding tasks. The proper choice depends a lot on how you use your tablet. Quote:
|
Re: A faster N810??
In the conditions we have (very low consumption when CPU idle),
high frequency wastes battery time when the CPU waits a lot, either when doing busy wait loop (which are not common, I guess) or when waiting for the memory, which I expect is much more common, considering the small low-associativity caches and slow RAM. So, I suggest running some tests before advising people to use this option. Recall that Nokia probably ran those tests, and selected to run the N800 at 333MHz, rather than 400MHz, when scaling was not available. This suggests to me that the frequency does have a significant affect. It is useful the remember that while power draw is proportional to the frequency, the OMAP also reduces voltage when running at lower frequency, and power draw is proportional to square of voltage, so it is possible that the CPU takes 2.5 as much seconds to do something at 166MHz, but might draw 10% of the power per second during this time. |
Re: A faster N810??
Quote:
:D |
All times are GMT. The time now is 23:09. |
vBulletin® Version 3.8.8