gschem very very slow

[arndt_t]

The current git head (20090805) version of gschem looks nice but its very very slow!
The 1.5.0 (version with guile 1.6, without cairo) is much more faster.
What could be the problem?

Env:
- gschem @ Sunray 2FS thin-client !?

[Peter TB Brett (peter-b)]

What graphics hardware are you using? Is gschem running locally, or on a remote host? If on a remote host, is the network being saturated? Is it being CPU-limited or IO limited?

What versions of GTK, Cairo, GLib and Guile have you compiled gEDA against?

(FWIW, on my system, the latest git version is as fast or possibly *slightly faster* than the pre-Cairo version. This is just my subjective perception, though.)

[Peter Clifton (pcjc2)]

Could be that the video drivers on your thin client aren't well optimised for cairo. These kind of things can make quite a big difference.

The main slow down is probably text drawing. That is set to change before 1.6.0 (to using "real" fonts - using the pango library).

Once the pango support is merged, I'd be interested to hear if your speed problems go away.

[arndt_t (arndt-t)]

@petercjclifton
This is possible i will ask for an update of my Sun SunRay 2FS.
I've not tested the currend GIT head, When i do this i give you a reply!

@peter-b
I run gschem remote on a linux RHEL4.0 system. My thin client has a RMI Alchemy Au1550 CPU and is served by a sunray server with solaris.
atk-1.9.1
cairo-1.2.6
gd-2.0.35
gettext-0.16
glib-2.12.13
gtk+-2.10.14
guile-1.8.7
pango-1.12.4

It may be much moch faster on a desktop computer or laptop with local graphic card.

[Peter TB Brett (peter-b)]

Now that 1.6.0 has been released, which uses "real" fonts, please test again and see if performance has been improved.

[Peter TB Brett (peter-b)]

Please test with 1.6.x or unstable head, and see if this problem is still present.

[Launchpad Janitor (janitor)]

[Expired for gEDA because there has been no activity for 60 days.]

[Peter TB Brett (peter-b)]

This bug relates to a very old version of gschem, and without any cooperation from the reporter, this bug will be impossible to debug or fix. Setting to "Invalid" -- if similar problems come up, a new (and hopefully more useful) bug should be filed.

[DJ Delorie (djdelorie)]

Note: I've found that the cairo version of gschem is very slow also.
We're sending a LOT of triangles to the X renderer.