[Gs-devel] Unicode CJK fonts and ghostscript/ghostview performance

Nelson H. F. Beebe beebe at math.utah.edu
Wed Nov 7 07:24:55 PST 2001

I just took Raph's pointer to


fetched the file, and launched acroread (4.05), xpdf (0.93), and gv
(ghostscript 7.03) to view it.  

I then pulled down the gv magnification menu, which I've customized to
go as high as 64, selected 64, and voila: the window went full screen,
and my workstation (a Sun UltraSPARC 170 running Solaris 2.8) locked
up for several minutes while the X server rendered CJK fonts!

I've reproduced this on a GNU/Linux box (Red Hat 6.2), and ran top
during the experiment: both X and gs hogged the CPU, though not for as
long as on my Sun.

Similar magnification in acroread had no such performance problem.
Perhaps it is because acroread has the most sensible magnifier of any
PDF viewer I've used: drag a rectangle around the region of interest,
and that region is magnified to fill the current window, but
preserving aspect ratio.  Presumably, the only thing that got rendered
at high resolution was the one CJK character that filled the window.

gv's magnification scheme is horrible: pull down a menu, guess at a
magnification factor, watch the region of interest vaporize from the
screen, and then try to find it again by dragging the image around.

As my experience today shows, it is even worse, because the entire
window gets rerendered at higher magnification, even if most of it
might never be viewed.

With acroread, I can use the hand icon to move the magnified image
around, and performance seems quite reasonable.


- Nelson H. F. Beebe                    Tel: +1 801 581 5254                  -
- Center for Scientific Computing       FAX: +1 801 585 1640, +1 801 581 4148 -
- University of Utah                    Internet e-mail: beebe at math.utah.edu  -
- Department of Mathematics, 322 INSCC      beebe at acm.org  beebe at computer.org -
- 155 S 1400 E RM 233                       beebe at ieee.org                    -
- Salt Lake City, UT 84112-0090, USA    URL: http://www.math.utah.edu/~beebe  -

More information about the gs-devel mailing list