xorg-x11-server-Xorg-1.7.0-1 is fast

[SlowJet]

With a dozen or so updates post rawhideand kernel 58, I just added these and wow, Gnome zips.

phonon-4.3.1-102.fc12.i686.rpm
phonon-backend-gstreamer-4.3.1-102.fc12.i686.rpm
phonon-backend-xine-4.3.1-102.fc12.i686.rpm
qt-4.5.2-23.fc12.i686.rpm
qt-x11-4.5.2-23.fc12.i686.rpm
xorg-x11-server-common-1.7.0-1.fc12.i686.rpm
xorg-x11-server-Xorg-1.7.0-1.fc12.i686.rpm

SJ

[leigh123linux]

Originally Posted by SlowJet

With a dozen or so updates post rawhideand kernel 58, I just added these and wow, Gnome zips.

phonon-4.3.1-102.fc12.i686.rpm
phonon-backend-gstreamer-4.3.1-102.fc12.i686.rpm
phonon-backend-xine-4.3.1-102.fc12.i686.rpm
qt-4.5.2-23.fc12.i686.rpm
qt-x11-4.5.2-23.fc12.i686.rpm
xorg-x11-server-common-1.7.0-1.fc12.i686.rpm
xorg-x11-server-Xorg-1.7.0-1.fc12.i686.rpm

SJ

Can you prove it's faster ?

[Hlingler]

xorg-x11-server-Xorg-1.7.0-1 is fast
With a dozen or so updates post rawhideand kernel 58, I just added these and wow, Gnome zips.

Huh. Wonder what's broken now....


V

[axet]

faster then 1.6.99? or faster then ever?

[SlowJet]

No, it is not any faster.
Something else must have done it. Maybe less debugging or less errors generated?
I noticed the change from Logn to Desktop.

Pent III 800 slow disk nouveua mx 32MB
Two boots with each version show no differnce.
From Bios to Plymonth text gone. 1M:33S
From Ply gone to Logon. 16S
Fron Logon to Gnome Desktop 33S

SJ

[hephasteus]

Probably temporarily stepped outside of linux's DON'T USE ANY RAM! IT'S A SIN!! LET'S be as small as we possibly can be and swap out all these processes. An annoying little habit that needs to be broken on X86_64. Reboot probably got everything into the sleeping coffins.

[SlowJet]

Originally Posted by hephasteus

Probably temporarily stepped outside of linux's DON'T USE ANY RAM! IT'S A SIN!! LET'S be as small as we possibly can be and swap out all these processes. An annoying little habit that needs to be broken on X86_64. Reboot probably got everything into the sleeping coffins.

http://kerneltrap.org/node/3000

No, that was not the issue.
But if you are so worried about the miss-understanding of swapiness, as hundreds of thousands of desktop users are, then it is very simple to fix.

If Desktop
If 1 and only 1 user loggin at a time (and maybe 2-4 inactive with memory to support them.)
echo 0 > /proc/sys/vm/swappiness
else
echo 100 > /proc/sys/vm/swappiness
else if sever with many users
echo 100 > /proc/sys/vm/swappiness
else
echo 60 > /proc/sys/vm/swappiness (which is the default now, used to be 90)

Results:

echo 100 > /proc/sys/vm/swappiness (The more data that is read and written, the more ide code pages will be swapped.)

echo 0 > /proc/sys/vm/swappiness (data and code fight over pages during low memory times. But code pages are not swapped out during idle times.)

echo 60 > /proc/sys/vm/swappiness (a compromised based on opposing sides that makes no sense at all.)

echo 90 > /proc/sys/vm/swappiness (A setting based on random personality of whomever had to initially set swapiness and make no sense at all.)

Your computer makes several billion decisions every second, 20 million Linux users can't make 1 in ten years.

SJ