speedup desktop response

[drr00t]

Hi guys

Somebody has test this: http://www.webupd8.org/2010/11/alternat ... patch.html?

shows nice tweak to have by default in Mint.

[JasonLG]

Hi guys

Somebody has test this: http://www.webupd8.org/2010/11/alternat ... patch.html?

shows nice tweak to have by default in Mint.

It's already included by default in the current spin of LMDE. Read the release announcement.
http://blog.linuxmint.com/?p=1604

Besides it only helps if programs are run from separate tty's. It has no effect on desktop apps unless you run each one from a separate terminal. Basically it's only useful if you plan on doing other stuff with your computer while you're doing CPU intensive tasks like compiling. So for the average n00b who does little to no compiling and such it's of little to no value. Any perceived everyday desktop speedup is purely a result of the placebo effect.

But if you really want or need the patch here is a link to a webupd8 post that has a script that will do it automatically.
http://www.webupd8.org/2010/11/script-to-automatically-apply-200-lines.html

[joutlan]

Nice....wonder if it helps me....most I do is running several apps within a VM, and 3 or 4 linux apps (well, MS Word and Excel via CrossOver) running at the same time.

[drr00t]

Hi, very nice. I work with Blender and C++, this help me a lot.

thanks.

[rivenathos]

I can tell a difference when using HandBrake.

[JasonLG]

I can tell a difference when using HandBrake.

Not unless you run it from a terminal.

The patch works something like this AFAIK.
The kernel scheduler tries to be as far as possible about dividing the CPU up between processes normally. What the patch does is group processes by tty. All desktop apps run from the same tty unless started in separate terminals so it doesn't do a darn thing for them unless they are started each from different terminals. Even then the effect is that if say without the patch program A starts 5 processes that each use 10% of the CPU(for a total of 50%) and program B starts 3 processes that each use 10% of the CPU(for a total of 30%) and program C starts 2 processes that each use 10%(for a total of 20%) altogether they use up 100% of your CPU. The patch groups the processes so that each program regardless of how many processes has an equal crack at the CPU in effect making the completely fair scheduler more fair on a per program basis. It doesn't actually speed anything up, It causes more responsiveness because it throttles back CPU hogs and lets other programs get a shot at the CPU.

[rivenathos]

Not unless you run it from a terminal.

Jason, you may be right. I was attributing the speed increase and subsequent faster encoding to the patch. However, in reality, it was the switch from LM10 Ubuntu-base to running pure Debian 6 Squeeze.

Never mind me, carry on....