Subject: Re: Heavy /dev/lpt use burdens system.
To: Brook Milligan <brook@biology.nmsu.edu>
From: Richard Rauch <rkr@rkr.kcnet.com>
List: port-i386
Date: 09/10/1999 18:22:56
Thanks to those who responded to my initial question.  lpa seems to be
much nicer to my system, though not ideal (see below).


>    > I have been trying the interruptless driver, but it goes incredibly
>    > slowly.  Like a line a minute or something.  I was thinking of
>    > switching to the interrupt driver, but now I'm not so sure that will
>    > help.
>    > 
>    > What am I doing wrong?
> 
>    What sort of printer?
> 
> Just an Epson dot matrix dinosaur.

My new printer replaces an old Epson dot matrix printer.  The old Epson
managed to get out about 1 page (rather than line) every few (4?)
minutes when taking Ghostscript output.  Slow, but nowhere near as slow as
you describe.  I think that I always drove it with the interrupt-based
lpt.  The printer had an 8K buffer; turning the buffer on or off didn't
seem to affect much of anything.  (The performance seemed to be limited by
the speed at which the data could be sent over the parallel port.)

(For another point of contrast: I get comparable performance on my new
laser printer when I generate the data to a file & cat the file to
/dev/lpa and, on the other hand, when I have GhostScript running as a
printer-filter (generating the data on-the-fly).  So, even though lpt
bogs my system down with the new printer, it doesn't adversely affect
overall printing speed.)

When dumping pre-computed printer data out the interruptless lpa, the X
mouse cursor is much less jumpy than when printing with lpt, but there is
still noticable jumpiness (maybe 10 or 20 positional updates per second?),
suggesting that perhaps even lpa isn't as nice as it could be.  (Or else
that the parallel port really isn't the best way to hook up a printer
requiring so much data per page?  *grin*)


  "I probably don't know what I'm talking about."  --rkr@rkr.kcnet.com