[Vis5d-users] Re: VIS5D problem with spikes in surface (land and water)

Dear Christopher,

I made the same experience with our SGI Onyx2 IR (Irix6.5). The bug has
somehow to do with the optimization of the routine proj.c.
If you recompile proj.c with the flag "-O1" instead of "-O2" or "-O3"
and then rebuild vis5d the bug should not occur.
Best Regards,
Michael

At 14.05.01 12:15, Champagne, Christopher wrote:
>Hi,
>
>I have installed VIS5D on an SGI machine (32 bit architecture, IRIX 6.5) 
>and have tested out the examples (LAMPS.v5d and SCHL.v5d) and have had 
>some problems with spikes in the surfaces of these pictures.
>
>I then used a utility from NCAR to reformat MM5 data into VIS5D format and 
>plotted the MM5 data in VIS5D.  The surface spike problem is even worse 
>with MM5 data as shown in the attachment (gif).  In the MM5 view the 
>picture is looking towards the Southeast US and FL.  The problem seems to 
>be worse over water than land with MM5, but I still do get random 
>occurrences of spikes on land too (especially near the edge).  Upon 
>further investigation it seems the spikes appear randomly, yet the height 
>of the spikes always seem to be half of the height of the VIS5D box used 
>in viewing and rotating the data.
>
>Has anyone ever encountered this problem before?  I am wondering if its an 
>installation problem, an Open GL problem, or a source code problem?
>
>Thanks for your time,
>
>Chris Champagne
>
>

_______________________________________________________
Michael Boettinger
Deutsches Klimarechenzentrum
Bundesstrasse 55
20146 Hamburg
Germany
email: boe...@dk...
phone: +49 40 411 73 344
fax: +49 40 411 73 270
http://www.dkrz.de/
_______________________________________________________