[jdaviestx]

Ok, this is strange. I recently upgraded (from source) from 2.6.18 to 2.6.24. Everything seems to work OK (finally), except... whenever I launch an application for the first time, it loads s l o w l y. After that, the same app launches painlessly (so, presumedly, it's being brought into a cache, made memory-resident, something).

I didn't observe this behavior in 2.6.18. Is anybody aware of some new 2.6.24 feature that might cause this type of behavior? Has this happened to anybody else?

Thanks.

[i92guboj]

Originally Posted by jdaviestx

Ok, this is strange. I recently upgraded (from source) from 2.6.18 to 2.6.24. Everything seems to work OK (finally), except... whenever I launch an application for the first time, it loads s l o w l y. After that, the same app launches painlessly (so, presumedly, it's being brought into a cache, made memory-resident, something).

I didn't observe this behavior in 2.6.18. Is anybody aware of some new 2.6.24 feature that might cause this type of behavior? Has this happened to anybody else?

Thanks.

For some people (including me) the CFS task (not i/o) scheduler, seems responsible for some responsiveness (interactivity) problems. But it shouldn't be related to your problem, if the only thing you noticed is longer startup times.

Maybe you disabled your chipset driver and so you are using non-dma transfers on your hd's? You only have to worry about that on IDE drives, so, if you use SATA that's not it. Or maybe it is just some placebo effect.

EDIT: Just in case, you might want to try recompiling your kernel after you turn this option off:

[ ] Fair group CPU scheduler

That's under the general settings.