> On Mon, Sep 08, 2003 at 05:08:52PM -0400, Nathan J. Williams wrote:
> > 
> > Have you measured it?
> 
> Itojun posted some numbers - if I read those correctly there is a 1000
> times slowdown in the RPC calls.  That is a rather large overhead even
> for an infreqent call.

	could anyone with lmbench do lat_rpc with the old code?

	the numbers i've posted (hoge.c) could be misleading, as (1) it only
	measures the algorithm itself, not the impact to RPC code as a whole,
	(2) it runs on userland so kernel-userland overhead like gettimeofday()
	call which makes it inaccurate.

itojun