Subject: Re: Benchmarking 10-base-T to 100-base-TX networks... (fwd)
From: Ron Chmara (ron@Opus1.COM)
Date: Sat May 06 2000 - 18:07:24 EDT
Spatch wrote:
> Don't know if anyone would care, but here is a benchmark of network
> thruput using 10Base-T and 100Base-TX networks comparing a lot of things
> (read the how to read this document).
> 5) Compare Mac thruput to PC thruput. This is by the way, the least fair
> test since the Mac is so much slower than the PC, at least in
> clock speed, and because the server side apps are written by
> different groups and will likely themselves cause different times
> to come out. This is why I used the ftp method of copying one
> file in Table 3. Both the Mac and PC use the same server FTP software.
And you used one heck of a slow Mac. Much of the 601 series had motherboard throughput
issues.
> Mac-OS 7.6.1 software configuration:
> AppleShare version: 3.6.5, Appletalk driver version: 60.1.2, Open
> Transport and TCP/IP version: 1.1.2, Speed Doubler 8.1.2 using faster
> network copy protocol.
This is interfering with (and occasionlly slowing down) your copies as well.
Speed doubler was great for appleatlk, with TCP/IP speed doubler peers/servers,
but now, ouch.
> If you must know how I am personally biased, I make no attempt to hide the
> fact that I prefer Mac's to Windows. If there was any unconscious bias
> introduced into this testing it would have been pro-mac, however, the
> numbers dictate otherwise. I did have to struggle with getting the Mac
> times that I did.
So fix 'em. Chuck "speed doubler", test with a Mac current to the PC you were
testing. You can get much higher speeds off of mac clients than that:
http://www.opus1.com/ron/asipstats.html
Speed doubler is designed for peer-peer networking, to bypass AppleshareIP,
no?
-Bop
-- Brought to you from iBop the iMac, a MacOS, Win95, LinuxPPC machine, which is currently in MacOS land. Your bopping may vary.
This archive was generated by hypermail 2b28 : Wed Jan 17 2001 - 14:30:37 EST