[rescue] When is 100 less than 10?

[Dieter Montanez]

I had once exactl the same problem, but with a SUN system.
All X stuff was very slow the rest was OK. The problem
turned to be the autonegotiation. Using "ndd" on the
solaris box I disabled "autoneg" and tried different 
speed/duplex combinations and 100HD did perform the best. 
The original switch was a cheap one, later I bought a 
CISCO 2590 and everything was peachy (using autoneg).
Is there something like "ndd" for the IRIX ?

regards

Dieter

-----Original Message-----
From: rescue-bounces at sunhelp.org [mailto:rescue-bounces at sunhelp.org]On
Behalf Of Sheldon T. Hall
Sent: Tuesday, April 26, 2005 5:52 PM
To: Sun Rescue List
Cc: Discussion of Large SGI Machines
Subject: RE: [rescue] When is 100 less than 10?


-----Original Message-----
From: Sheldon T. Hall [mailto:shel at artell.net] 
Sent: Tuesday, April 26, 2005 2:47 PM
To: 'The Rescue List'
Cc: Discussion of Large SGI Machines (big-iron at sgihelp.org)
Subject: RE: [rescue] When is 100 less than 10?


OK, this gets weirder.

I've tried switching switches, cables, and ports on the Challenge's vfe
board, and it's all the same.

However, what I have discovered is that, while X applications are bog slow
over the 100baseT link, slower than over 10baseT, normal apps are faster.
Using vfe0/100baseT, nedit or fm take 15 to 20 seconds to open a window on
my laptop, but vi or pico have their new window open before I can get my
pinkie off the return key.  Using et0/10baseT, nedit or fm have a window
open in about 3 seconds.

In addition, using vfe0/100baseT, nedit gets slower and slower, until it
becomes unusable.  Even closing the window takes 20 seconds at that point.
It works fine under et0/10baseT, of course.

So ... my question evolves:  Why are X apps over vfe0 slower than over et0,
while everything else is faster?

-Shel
_______________________________________________
rescue list - http://www.sunhelp.org/mailman/listinfo/rescue