*BSD News Article 34101


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msuinfo!agate!howland.reston.ans.net!spool.mu.edu!torn!newshost.uwo.ca!home-pc1.business.uwo.ca!MARK
From: MARK@ardsley.business.uwo.ca (Mark_Bramwell)
Newsgroups: comp.os.386bsd.bugs
Subject: Re: Final word on 1.1.5.1 and Dhrystone. WRONG Answer
Date: Wed, 10 Aug 1994 04:39:34 GMT
Organization: Western Business School
Lines: 46
Message-ID: <MARK.1080.2E485A06@ardsley.business.uwo.ca>
References: <MARK.1057.2E352DF5@novell.business.uwo.ca> <3170dq$1ko@ug1.plk.af.mil> <miff.776242584@apanix.apana.org.au>
NNTP-Posting-Host: home-pc1.business.uwo.ca

In article <miff.776242584@apanix.apana.org.au> miff@apanix.apana.org.au (Michael Smith) writes:
>From: miff@apanix.apana.org.au (Michael Smith)
>Subject: Final word on 1.1.5.1 and Dhrystone. (was re: <blah>)
>Date: 7 Aug 94 06:56:24 GMT

>The deal with the dhrystone change is simple :

>in 1.1 and lower, HZ was 60, which is fairly normal for *nix systems.
>With 1.1.5.1 it was changed to 128 (as told to me by newton@cleese).

>As previously reported, some versions of Dhrystone assume HZ is 60, others
>make no assumption and require you to specify a value.

>Dhrystone 2.1 allows you to do this by specifying a definition of HZ when
>you compile, ie -DHZ=60 or -DHZ=128 as appropriate.

>Application speed does not appear to have changed significantly between the 
>two versions 8).



I have been seeing many replies from my first question and all have been the 
same and all have been wrong.  Yes HZ had been changed from one kernel to 
the next, however, I did say the code was recompiled.  Also, the code uses 
the defined value for HZ which is in the include files.  In other words, the 
recompile should have picked it up.  Saying that I should use a better 
benchmark is not an answer.


Note:  I don't feel the machine is any faster or slower.


This is my guess....  Since I was using SUP to download the newer kernels, I 
had a mixure of old (1.1) and new (1.1.5beta).  Most of the libs and other 
things were still based on the old stuff, whereas the kernel was new.  This 
probably caused false (high) readings with the older kernel but valid 
readings with the all new kernel.  

I hope that made sense!?


=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Mark Bramwell, VE3PZR                Located in sunny London, Ontario

Internet: Mark@ARDSLEY.business.uwo.ca  IP Address: 129.100.22.33
  Packet:  VE3PZR @ VE3GYQ               UWO Phone: (519) 661-3714