Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!usc!cs.utexas.edu!uunet!lll-winken!scooter!neoucom!wtm
From: wtm@neoucom.UUCP (Bill Mayhew)
Newsgroups: comp.sys.att
Subject: Re: cu on a 3b1 is SLOW
Summary: Try kermit,  pcomm, Honey DanBer, etc
Message-ID: <1696@neoucom.UUCP>
Date: 4 Aug 89 02:16:19 GMT
References: <18818@mimsy.UUCP> <4800003@spdyne>
Sender: wfd@neoucom.UUCP
Organization: Northeastern Ohio Universities College of Medicine
Lines: 29


Part of the limitations of throughput with the stock version of cu
supplied with [version 3.51 software of] the 3b1/Unix PC is that
the main loop only reads and puts one character to the screen at a
time; which is horribly slow, as you might guess.

A better way to work is to write a block of characters to the
console at once.  This requires a little more work, as the input
routine shoud set a timer that forces it to dump a short block if
the requisite amount of characters haven't been received in n
milliseconds.

By the way, the 3b1 console makes almost 4800 baud effective rate
when catting a file from the local winchester, so you'll be
somewhat constrained by that.  (Kinda hard to read faster than
thant anyway!)

There isn't much that can be done in the way of trying to patch the
stock cu.  There are a number of publicly available terminal
emulators that do work, however.  I use C-kermit 4E(070) dated Jan
29, 1988 on my 3b1 here.  I use a PC7300 termcap on the vax I talk
with and have no problems.  Pcomm was posted to the net a while
back, and it looks very much like procomm of the msdos world; I
have not used pcomm, but have heard good comments.  You can also
try to get your mits on the illusory Honey DanBer uucp package, as
it has a fixed cu with it.


Bill     wtm@impulse.UUCP   or   wtm@neoucom.UUCP