Re: Double Precision in 2.0?
Andrew Shein (Andrew.Shein++at++orac.boston.sgi.com)
Thu, 8 Feb 1996 21:32:22 -0500
On Feb 8, 4:37pm, Gary Williams wrote:
> Subject: Double Precision in 2.0?
> I've been continuing work on a virtual Solar System Modeler which accurately
> propagates the orbits of the planets, moons, man-made satellites, etc. We're
> experiencing some difficulties with jitter in the moons and satellites in the
> outer planet regions. The coordinate system is based on the Earth's location
> so the numbers get pretty large. The calcuations must be precise and I
suspect
> our problem may be due to the mixing of doubles and floats (pfVec3, etc) in
> Performer 1.2. I've been told Performer 2.0 was to support double precision
in
> its internal data structures. However, it appeared to me the pfVec3 and
related
> structures are still only floats (I looked at the pfLinMath.h file). I've
> briefly looked at the FAQ page and the manual but haven't found any helpful
> information. Would someone please let me know if Peformer 2.0 does support
> doubles in its data structures and an appropriate reference for reading up on
> this. Thanks for any help you can provide.
>
> Gary
> ******************************************************
> GARY E. WILLIAMS, Capt, USAF
> Air Force Institute of Technology
> MS Student in Computer Systems: Modeling & Simulation
>
> http://www.afit.af.mil/Schools/EN/ENG/LABS/GRAPHICS/people/gewillia/index.html
> gewillia++at++afit.af.mil (513) 255-3636 Ext. 6036
>
>
>
>-- End of excerpt from Gary Williams
If I remember correctly the GEs work in floats, so
even if you do the positions as doubles, they are changed to floats.
I had the same problem with a gl app a few years ago.
You might try changing the coordinate system (dynamicly) to the closer
objects, that way the jitter will not be as noticable on the
far away objects.
Andy
--
Andrew Shein SE Stout email: ashein++at++boston.sgi.com
Silicon Graphics Inc. phone: (508) 562 - 4800
1 Cabot Road fax: (508) 562 - 4755
Hudson, MA 01749 vmail: 59688
This archive was generated by hypermail 2.0b2
on Mon Aug 10 1998 - 17:52:23 PDT