comp.lang.idl-pvwave archive
Messages from Usenet group comp.lang.idl-pvwave, compiled by Paulo Penteado
Home » Public Forums » archive » Re: Object graphics under Linux: are they supposed to be that slow?
Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend  		Return to the default flat view Create a new topic Submit Reply
Re: Object graphics under Linux: are they supposed to be that slow? [message #29032 is a reply to message #29029] Tue, 29 January 2002 19:00 Go to previous messageGo to previous message
Mark Hadfield is currently offline  Mark Hadfield
Messages: 783
Registered: May 1995
Senior Member
Several people have asked me for a copy of the program on which I
based my assessment that IDL object graphics is considerably slower in
Linux than in Windows. I haven't made it available yet, because it is
entangled deeply in my IDL libraries; it doesn't make a very good
benchmark anyway because it is event-driven and it's hard to include
the timing functions. What I do have is an object-graphics benchmark
program called TIME_TEST_GR2. This was originally written by Randall
Frank and copyright is held by RSI. I am making it available by kind
permission of Randall Frank and Karl Schultz:

http://katipo.niwa.co.nz/~hadfield/gust/software/misc/time_t est_gr2.pro

This version is functionally identical to the original but prints out a
little more information about the window and device properties.

Perhaps TIME_TEST_GR2 could be part of JD's revised benchmark suite?
However it does have the drawback that it takes quite a while to run
(~ 10 minutes) and you can't do much else on the system while it's
running.

I have also produced 3 files of TIME_TEST_GR2 output generated on my
system:

http://katipo.niwa.co.nz/~hadfield/gust/software/misc/ttgr2_ Windows.txt

http://katipo.niwa.co.nz/~hadfield/gust/software/misc/ttgr2_ Windows_Renderer
_1.txt
http://katipo.niwa.co.nz/~hadfield/gust/software/misc/ttgr2_ Linux.txt

The configurations are:

* IDL 5.5 on Windows 2000 using RENDERER=0 (hardware)
* IDL 5.5 on Windows 2000 using RENDERER=1 (software)
* IDL 5.5 on Linux. This uses RENDERER=0 but, as is obvious from
the DeviceInfo string, the rendering is carried out by the Mesa
software library and does not access any hardware acceleration

The geometric-mean elapsed time figure provides a rough ranking of the
configurations:

Windows RENDERER=0 4.58 s
Windows RENDERER=1 3.11 s
Linux 5.55 s

The surprise here is that Windows is slower with RENDERER=0. So much
for hardware acceleration! This suggests my graphics controller is no
ball of fire. Linux lags behind both the Windows configurations,
though not by as much as my original assessment suggested.

I have checked that the Windows and Linux configurations are as
similar as possible. The screen dimensions (1280 x 1024) and colour
depth (16 bits) are the same in both. I think I may have already
mentioned that the PC's graphics controller (Intel 810) has only 4 MB
on-board memory (at least that's what the Windows Display properties
applet tells me).

I intend to investigate the effect of reducing the number of pixels on
the screen. I also intend to investigate the effect of grovelling to
our IT people for some better graphics hardware.

The detailed results to TIME_TEST_GR2 are interesting. Windows with
RENDERER=0 is notably slow in the instancing test. Linux is notably
slow in the polygon test.

---
Mark Hadfield
m.hadfield@niwa.co.nz http://katipo.niwa.co.nz/~hadfield
National Institute for Water and Atmospheric Research

Report message to a moderator

[Message index]
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: Integration of a complex function
Next Topic: Re: Subject : locks, semaphores, and such

-=] Back to Top [=-
[ Syndicate this forum (XML) ] [ RSS ] [ PDF ]

Current Time: Thu Oct 09 22:53:26 PDT 2025

Total time taken to generate the page: 0.95789 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.7.
Copyright ©2001-2013 FUDforum Bulletin Board Software