File: README.performance

package info (click to toggle)
printbill 4.1.2-1
  • links: PTS
  • area: main
  • in suites: sarge
  • size: 660 kB
  • ctags: 151
  • sloc: perl: 5,488; ansic: 271; sh: 240; makefile: 108
file content (24 lines) | stat: -rw-r--r-- 1,327 bytes parent folder | download
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
* Try to make tmpfs live on the same filesystem as your spool directory -
e.g. if /var is a partition, and /var/spool/lpd/* are your spool dirs, it's
a really good idea to make /var/tmp the tmpfs. It makes a big difference -
particularly noticable when printing larger files. If you have everything on
one big root partition, it will be OK.

* If memory is tight, recompile gs to use a minimal set of drivers (just png
should suffice for billing). I have no figures to say how big the
improvement is though, so results may not be spectacularly better. Optimise
for your particular CPU (e.g. using gcc 3.0.x you can specify
-march=i686 or -march=athlon) - it should help.

* Heavy printing environments may wish to try printbill_lazybill instead of
printbill_scheduler - but make sure you have plenty of disk space as the
billing tasks could build up quite quickly.

* If you have a really busy print queue you might want to try using MOSIX
transparent clustering / load balancing to distribute the load. I have no
idea if this would actually work - the overhead in shifting the PS files
around may be too great. If it works for you please tell me. SMP is best.

* Printing from Windows usually works best if you use Adobe's generic
PostScript driver. Apparently it is somewhat faster than many of the
printer-specific drivers.