It might be a fairer comparison to rime a cross compilation
	of a pmax kernel on an alpha... I seem to remember someone
	saying the optimiser on the alpha did a lot more work than	
	most..
	(Unless of course the actual speed metic really is just how long
	you have to wait before you can boot your new kernel :)

		David/absolute

Definition of the day:
'Momentarily': _For_ an moment... not _in_ a moment...
(It would be preferable if more of the world could grasp the above)

On Wed, 4 Feb 1998, Toru Nishimura wrote:

> Date: Wed, 4 Feb 1998 17:46:20 +0900 (JST)
> From: Toru Nishimura <nisimura@is.aist-nara.ac.jp>
> To: port-pmax@NetBSD.ORG
> Subject: Re: kernel compile 
> 
> In article <199802020633.RAA08407@balrog.supp.cpr.itg.telecom.com.au>
> 	simonb@telstra.com.au wrote
> 
> >> As Ted says, building a kernel from scratch takes about an hour.
> > 
> > Hmm, about 25 minutes on the /260 :)
> 
> Let me repeat what Emanual said recently.
> 
> 	heaven. it must be heaven. :-))
> 
> My DS5000/33 with 40MB memory compiles GENERIC about 50 minutes.
> 
> It's somehow interesting to compare NetBSD/pmax with NetBSD/alpha.
> DEC3000/300, a natural descender of Personal DECstation, has 150MHz
> 21064 CPU with 64MB memory, and it compiles TC-only configuration
> about 40 minutes.  Even taking the fact NetBSD/alpha is larger and
> complicated into counts, 60MHz R4400 seems archieving very good
> performance.  I know 200MHz 21064 machine DEC3000/800 with 256MB
> compiles TC-only configuration about 20 minutes. 
> 
> I'd have the R4000 upgrade module. :->
> 
> Tohru Nishimura
> Nara Institute of Science and Technology
> 
> 
>