On Thursday 15 May 2008, Mick wrote: > > You'll probably get better results with X by selecting a suitable > > process scheduler and configuring your HZ to 1000 > > Now, this I have noticed making a difference. Not all schedulers are > born the same. I have found that (the current version of) CFQ is > better than others. > > As a matter of interest, I remember reading somewhere that squeezing > 1000Hz out of an old machine may have the opposite effect to that > intended. Is this pub talk, or have you experienced something that > confirms this?
No, it's not just pub talk. The trick is to look closely at what is happening and why. The HZ value indicates how often the kernel should "tick", which is a timing signal. The tick consumes resources of course, but has the benefit of accurate timing signals. Modern machines can cope with this nicely, they are fast enough. Older machines, in combination with the kind of software we run these days, can't cope with this amount of activity, and the whole system slows down. The problem is very dynamic and subject to many variables so there is no single one-shot solution. The answer to what to do very much starts with "It depends" This is why Con first started working on process schedulers, looking for an algorithm the kernel could use to adapt to these cases and still be responsive on the desktop. The man with the numbers to back it all up is Ingo Molnar. You can read the lkml archives from a few months back when CFQ was going through heavy development to get an idea of how tricky this can really get -- Alan McKinnon alan dot mckinnon at gmail dot com -- gentoo-user@lists.gentoo.org mailing list
