Re: [patch] swap-speedup-2.4.3-B3 (fwd)

[Mike Galbraith]

On Wed, 25 Apr 2001, Marcelo Tosatti wrote:

> On Tue, 24 Apr 2001, Linus Torvalds wrote:
>
> > Basically, I don't want to mix synchronous and asynchronous
> > interfaces. Everything should be asynchronous by default, and waiting
> > should be explicit.
>
> The following patch changes all swap IO functions to be asynchronous by
> default and changes the callers to wait when needed (except
> rw_swap_page_base which will block writers if there are too many in flight
> swap IOs).
>
> Ingo's find_get_swapcache_page() does not wait on locked pages anymore,
> which is now up to the callers.
>
> time make -j32 test with 4 CPUs machine, 128M ram and 128M swap:
>
> pre5
> ----
> 228.04user 28.14system 5:16.52elapsed 80%CPU (0avgtext+0avgdata
> 0maxresident)k
> 0inputs+0outputs (525113major+678617minor)pagefaults 0swaps
>
> pre5 + attached patch
> --------------------
> 227.18user 25.49system 3:40.53elapsed 114%CPU (0avgtext+0avgdata
> 0maxresident)k
> 0inputs+0outputs (495387major+644924minor)pagefaults 0swaps
>
>
> Comments?

More of a question.  Neither Ingo's nor your patch makes any difference
on my UP box (128mb PIII/500) doing make -j30.  It is taking me 11 1/2
minutes to do this test (that's horrible).  Any idea why?

(I can get it to under 9 with MUCH extremely ugly tinkering.  I've done
this enough to know that I _should_ be able to do 8 1/2 minutes ~easily)

	-Mike