Subject: Re: Compiler timings on varous MVII NetBSDs etc.
To: NetBSD Bob <nbsdbob@weedcon1.cropsci.ncsu.edu>
From: Lord Isildur <mrfusion@umbar.vaxpower.org>
List: port-vax
Date: 01/24/2001 15:07:57
the best board order is to put things with small buffrs and/or high 
interrupt rates closer to the processor. ethernet interfaces, async 
boards, TK controllers, put them close. disks go last. disk controllers 
tend to have buffers and while they might still dominate the total 
_traffic_ theyre not blotting out the other devices' interrupts. 
still, since what youre doing right now is all disk traffic, this shouldnt
havemuch effect. 
isildur

On Wed, 24 Jan 2001, NetBSD Bob wrote:

> > > I tried and it died.  perl5 would not compile on the MVII boxes.
> > 
> >  5:04 kurkku ~ %% perl
> > print "works fine.\n";
> > ^D
> > works fine.
> > 
> > Did this on a 13MB MicroVAX II, and yes, that is perl5 which I
> > compiled myself.
> 
> How?  My perl make was still making its makefiles 36 hours later!
> 
> What is different about your machine or my machine?  Something is
> making mine slow as molasses in january.
> 
> KA630
> SQ706A
> 13mb ram (forget the numbers right off)
> 4 x 1 gig scsi drives.
> TKQ50
> DEQNA
> 
> (not necessarily in that order, but, speaking of order, IFF one or
> more of the boards was not set right, that could slow things down?)
> 
> What are the boards and board ordering in your machine?
> 
> Thanks
> 
> Bob
>