Izumi Tsutsui wrote:

> In <200010271311.e9RDBlc05036@mirage.ceres.dti.ne.jp>
> I wrote:
> 
> > I tried to make a GENERIC3X kernel on both kernels which use
> > malloc/free and pool_get/pool_put:
>  :
> > I'll also try on CATS, macppc and newsmips.
> 
> I've tested on newsmips. "time make" on building GENERIC said:
> 
> malloc: 10770.145u 1191.785s 3:27:29.50 96.0%    0+0k 1064+14844io 1176pf+0w
> pool:   10556.347u 1184.212s 3:24:11.53 95.8%    0+0k 1074+14838io 1176pf+0w

Using Izumi's patches on a 60MHz R4400 (DECstation 5000/260) I saw a
slight _decrease_ in a single kernel build benchmark (building a pmax
almost-GENERIC kernel):

malloc: 2741.966u 422.555s 54:25.59 96.9%       0+0k 2528+20581io 1535pf+0w
pool:   2781.509u 423.518s 55:04.25 96.9%       0+0k 2433+20654io 1330pf+0w

This was done in single user using the same sequence of commands each
time (mount needed filesystems, cd, make).  I'll run the same thing
again to see if there is any difference a second time around.

> I think the results are notable. If there are no objections,
> I'll commit the changes for mips/pmap.c.

Simon.
--
Simon Burge                            <simonb@wasabisystems.com>
NetBSD Sales, Support and Service:  http://www.wasabisystems.com/