From: hooft@fys.ruu.nl (Rob Hooft) Subject: Re: cp foo.graphics /dev/lp1 is slooooooooow. ANSWER! Date: 28 Dec 1992 15:30:37 GMT
jwiegand@moe.eng.temple.edu (James Wiegand) writes:
>johnsonm@lars.acc-admin.stolaf.edu (Michael K. Johnson) writes:
>>bernd@iamk4526.mathematik.uni-karlsruhe.de (Bernd Wiebelt) writes:
>>
>> This has been asked before, but not yet answered.
>>
>>BZZZZZZZZZZZZT! Wrong. I have posted twice about this already.
>Actually, the answer is quite simple and no secret.
>Increase the value of LP_INIT_CHAR until you are happy.
>I have:
>#define LP_INIT_CHAR 250000
If you have these problems, please start by trying LP_INIT_CHAR 500,
and if necessary go even higher. Furthermore: at the moment it works,
mail the number to MKJ, so that he knows the results for different
printers!
>in include/linux/lp.h and get the same throughput as before. Otherwise,
>I was getting one line of graphics every 20 minutes. Pretty sh*tty!
The problem is that your printer is so slow in reacting, that the
driver thinks its buffer must be full. My 7 year old EPSON reacts
after 210 tries, so it is reasonable to assume that everything else is
faster than that, and certainly faster than 250! Well, it seems this
is not true.
Please note that it doesn't only have advantages to make the number as
high as you can, because when the printer buffer is really full a lot
of CPU time will still be spent trying to print one more character.
If you have a very fast laserprinter you may even want to lower the
number: that's what the new tunelp program will be able to do on a
running system!
>jim
>When in doubt, there is always the source code!
True.
-- Rob W.W. Hooft, Department of crystal and structural chemistry Bijvoet Center for Biomolecular Research, University of Utrecht The Netherlands ===== hooft@chem.ruu.nl (hooft@hutruu54.bitnet) ====Use a Real Operating System on your 386: Linux is FREE!====