Jean-Francois Richard 2275 (jfr++at++cae.ca)
Wed, 13 Oct 93 19:59:03 EDT
I am currently developing a Performer-based program to display
very large visual databases (too large to be loaded all at once).
My current solution is to divide the DB in a number of small tiles
and load those tiles as required by the current viewpoint.
The problem is that doing that in the main loop (or in the APP
process when using a multi-processing approach) creates ugly CPU
and I/O peaks whenever tiles are loaded and deleted. I need a smooth
update rate and can't afford to have these peaks.
I was planning on creating a forked process that would run at a low
priority and would load the tiles for the main process, attempting to
"look ahead" to figure out what was going to be needed soon.
Unfortunately, it seems that you cannot make Performer calls from
processes not started by Performer itself, which makes it impossible
to create or delete pfNodes from a forked process.
I'm stuck.
This looks to me like it should be a fairly common problem for
people with large amounts of data to display. Can anybody out
there think of anything to get around this problem?
--------------------------------------------------------------
J.F. Richard, jfr++at++cae.ca
CAE Electronics
Montreal
--------------------------------------------------------------
This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:50:03 PDT