POV-Ray : Newsgroups : povray.unix : POV-Ray 3.50c for Solaris Intel & UltraSparc at www.blastwave.org : Re: POV-Ray 3.50c for Solaris Intel & UltraSparc at www.blastwave.org Server Time
23 Dec 2024 12:45:17 EST (-0500)
  Re: POV-Ray 3.50c for Solaris Intel & UltraSparc at www.blastwave.org  
From: Dennis Clarke
Date: 18 Apr 2003 23:15:30
Message: <Pine.GSO.4.53.0304182304420.13585@blastwave>
On Fri, 18 Apr 2003, Thorsten Froehlich wrote:
>In article <Pine.GSO.4.53.0304180646450.11053@blastwave> , Dennis Clarke
><dcl### [at] blastwaveorg>  wrote:
>
>>    can this be reported at http://www.blastwave.org/bugtrack/  ??

I just read ( with great interest ) the fol :

// WARNING: The default uses POV-Ray's own tricks which only work if
// "float" is a 32 bit IEEE 754 floating point number!  If your platform
// does not use 32 bit IEEE 754 floating point numbers, radiosity will
// be broken!!!  If you have this problem, your only other choice is to
// use an ISO C99 standard revision compatible compiler and library:
//
// Define this to 1 to use ISO C99 functions logbf, logb and copysign.
// Define this to 2 to use ISO C99 functions ilogbf, ilogb and copysign.

Which causes me to wonder if the appropiate strict IEEE floating point rules
can be enforced on the Sun Forte compiler to ensure that things are handled in
a 32-bit way.  I really would need to brush up on the IEEE floating point
format to completely grok the implications here but I certainly understand the
issues of number conversion problems when bits get lost, tossed, shifted or
otherwise clearly misunderstood.  I was once a programmer in the bad old days
when pascal was considered a perfectly reasonable language to do systems
programming in even though it completely lacked the necessary tools to go from
floating point to long integers and vice versa.  Or maybe it was a long double
sin() or cos() function problem?  Can't remember .. that was back in 1985 and
the OS was CP6 on a Honeywell mainframe which had its entire OS written in
pascal and Fortran.  Regardless .. I get the concept of 32/64 bit float
problems with crystal clarity.

>
>No, because it has been fixed in 3.51, which is "ready" as far as the source
>code is concerned, but other issues currently prevent a release. It is
>expected in a few weeks nevertheless.

  hey!  any chance of getting a pre-release copy that I can work on now and
then I can get the diffs later?

  I am looking forward to the day when POV-Ray can harness multiple procs with
POSIX compliant threads ( I hope ) and then interesting animations will be
possible in near real-time.  All I need is a mainframe class system with a
terabyte of RAM and I will be happy.  I do have access to a Beowulf cluster
but it is busy doing profitable work and people would frown on POV-Ray running
at some low nice level :)


>
>    Thorsten

 Thank you very much ... you have been very helpful over the past few weeks.

 Dennis Clarke
 dcl### [at] blastwaveorg


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.