On Tue, 31 Oct 2000, Ken Wellsch wrote:

> Anders Magnusson wrote:
[...]
> > These are the two interesting lines. Well, what's left is for someone
> > to really profile the IP stack to find the problem...
> 
>            -------Sequential Output-------- ---Sequential Input-- --Random--
>            -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
> Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU  /sec %CPU
> 1.4.2   20    66 99.1   867 94.1   359 67.0    73 99.2   630 42.1  27.9 29.4
> 1.5_B   20    65 99.8   715 98.4   308 69.4    66 99.4   566 52.5  25.5 32.1
> 
> The disk benchmark also showed a similar drop in performance.
> 
> Might this be something more low-level?  I'm not sure what alas.

Time to check the code generated by gcc perhaps?
Have someone checked the performance of a simple user application?

	Johnny

Johnny Billquist                  || "I'm on a bus
                                  ||  on a psychedelic trip
email: bqt@update.uu.se           ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol