🚀 go-pugleaf

In article <1992Dec15.162702.16600@walter.bellcore.com> ngai@nova.bellcore.com (John Ngai) writes:
>
>In article <1992Dec15.141953.21972@dvorak.amd.com>, tdbear@dvorak.amd.com (Thomas D. Barrett) writes:
>|> In article <8f9__KS00WB5Qu_ogq@andrew.cmu.edu> fj05+@andrew.cmu.edu (Faisal Nameer Jawdat) writes:
>I watched this thread for a while, but am surprised no one points out an
>important reason of why you don't want to replace a DX50 with the DX2/66.
>The question one should ask is what performance do I gain? For all I have read
>and heard, the two chips delivers roughly the same *average* performance in
>most real applications. After you sum up the balance of on-chip and off-chip
>activities, it is simply about the same. Unless you are doing some
>computationally intensive task that are small enough to reside entirely in the
>on-chip cache, it is simply not worth it to swap.

Well here is a test you can try for yourself at home.

I selected a moderately large file and tried:

	time compress < bigfile > /dev/null

on two machines here. One of which has a DX-50 and one of which has a DX2-66.
In other respects the machines are fairly similiar. EISA based, moderately
fast SCSI based disks with 1742 controllers, 32MB of memory, both running 
SCO 3.2.4 etc.

The DX2-66 was about 1.21 times faster than the DX-50 when comparing the 
user time. 

I consider this to be fairly worthwhile on this machine. From my accounting
reports I see that compress recorded 46 CPU minutes over the last day. This was
slightly over 10% of total for the day. The only larger total was for uucico 
which was higher but spread over a much longer real time.

-- 
Stuart Lynne <sl@wimsey.com> ......................... UNIX Facsimile Software
Wimsey Information Technologies ................... moderator biz.sco.binaries
uucp login:nuucp passwd:nuucp .................... ftp.wimsey.com:~ftp/ls-lR.Z
PD Software for SCO UNIX .................... ftp.wimsey.com:~ftp/pub/wimseypd