[855] in testers
Re: Slow RT's
daemon@ATHENA.MIT.EDU (daemon@ATHENA.MIT.EDU)
Wed May 23 15:05:13 1990
From: cfields@ATHENA.MIT.EDU
Date: Wed, 23 May 90 15:04:25 -0400
To: testers@MIT.EDU
I tried setting maxtextcache to -1.
My quick conclusion: that wasn't it; my login was still very slow.
The rest is just FYI; don't read it unless you're interested.
Here are timings:
Kernel: 6.4R 7.0D 7.0D/patch
---- ---- ----------
Run: 1 2 1 2 1 2
---------- ---------- ----------
(a.) 34 29 47(45) 28 35 33
(b.) 34 26 45(41) 39 42 40
(c.) 39 35 60(67) 51 68 42
Units are in minutes. :)
Notes:
The 6.4R kernel and disk are running on top of 7.0 AFS packs.
There are two runs for each timing. The first runs happened directly
after a reboot before the packs were "detached"; the second series of
runs were done after the first, after a logout and ^P from XLogin to
return to console. They also started before the packs were "detached."
(a.) is the time from "hit any key to start" to the XLogin window
popping up
(b.) is the time from pressing return for the password until I got
a cshrc prompt. My .xsession file contains only a line to start
up an xterm.
(c.) is the time from sourcing my .startup.X file until I get a prompt
back. Once again, that file contains:
setenv DISPLAY unix:0
xrdb .Xresources
xset s blank
if (`machtype -c` == MVAX-II) xset m 25 3
if (`machtype` == rt) xset b 50 1000 50
if (`machtype` == rt) xmodmap .map
/mit/x11r4/`machtype`bin/twm &
xset +fp /mit/x11r4/`machtype`lib/X11/fonts/misc
xbiff -file /tmp/xbiff.cfields -update 10 -geometry +3-3 &
/usr/etc/zwgc
znol | zwrite cfields &
The numbers aren't that great. Probably +/- 5 seconds. The most important
numbers I think are for run one. Numbers in parenthesis are from the
original timings I reported. Others ar all new.
Craig
