> Depends - if you're not using X then you won't mind X getting swapped
> (running emerge from ssh/etc).  If you are using X then chance are it
> won't get swapped in the first place.  If the issue is gcc using all ram
> for itself, then the presence/absence of tmpfs probably won't make much
> difference - gcc still has to run.
>
> I don't see why swap should be any better/worse than any other form of
> disk access.  If anything it is superior as it allows the kernel to
> manage it like any other form of RAM, and the kernel isn't forced to
> flush it out to disk within some time (writes to a filesystem are forced
> to sync within some time to prevent data loss - this hurts performance
> and is totally unnecessary for temporary build files that will get
> deleted when you re-run emerge anyway).
>
> Even if you don't have a tmpfs writing to disk will tend to drive unused
> ram into swap - the system will swap idle memory to make room for
> cache/buffers - where the recently-written files will reside in ram.
>
> Now, in an extreme case where you have less RAM than the resident size
> of the apps you have running then swap will be horrible - but that is
> because you're continuously swapping in and out.  And I doubt a tmpfs
> will make much difference either way.
>
> Now, if somebody has empirical data I'll certainly pay attention, or at
> least a lot of expertise.  However, I wouldn't jump to the conclusion
> that swap is worse than ordinary disk writes - linux manages swap fairly
> well all things considered.  If you don't like how it is being managed
> there are kernel settings that can be used to tweak it (swappiness, etc).

Yes you are completely right, but that is not the point I wanted to make. I 
just think avoiding any unnecessary disk access is the best solution as long 
there will be the bottleneck.

Rgds
Bernhard
-- 
[EMAIL PROTECTED] mailing list

Reply via email to