 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Would it be possible/easy to compile a version of Povray which uses 16
bit floats?
-Shay
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
In article <3eb15e33@news.povray.org> , "Shay" <sah### [at] simcoparts com> wrote:
> Would it be possible/easy to compile a version of Povray which uses 16
> bit floats?
There is no 16 bit float format supported by any existing hardware, so what
would it be useful for? Are you even sure you are talking about floats and
not integers?
Thorsten
____________________________________________________
Thorsten Froehlich, Duisburg, Germany
e-mail: tho### [at] trfde
Visit POV-Ray on the web: http://mac.povray.org
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
In article <3eb15e33@news.povray.org>, "Shay" <sah### [at] simcopartscom>
wrote:
> Would it be possible/easy to compile a version of Povray which uses 16
> bit floats?
I don't think you quite know what you're asking. POV-Ray uses double
precision floats for a reason, 16 bits would not have anywhere close to
the precision needed. It might be useful for storing large dynamic range
images on disk, but would be slower for calculations than the native
floating point types. There are a few places where 16-bit fixed point
arithmetic might be useful (and probably several places where fixed
point arithmetic of some kind is actually used), but it isn't something
you would change all of POV-Ray to use.
Why would you want a version that used 16 bit floats?
--
Christopher James Huff <cja### [at] earthlinknet>
http://home.earthlink.net/~cjameshuff/
POV-Ray TAG: chr### [at] tagpovrayorg
http://tag.povray.org/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Thorsten Froehlich" <tho### [at] trfde> wrote in message
news:3eb165da$1@news.povray.org...
|
| There is no 16 bit float format supported by any existing
| hardware, so what would it be useful for? Are you even sure
| you are talking about floats and not integers?
|
I see. I was not aware of this. Thank you.
-Shay
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Thorsten Froehlich" <tho### [at] trfde> wrote in message
news:3eb165da$1@news.povray.org...
> In article <3eb15e33@news.povray.org> , "Shay" <sah### [at] simcopartscom> wrote:
>
> > Would it be possible/easy to compile a version of Povray which uses 16
> > bit floats?
>
> There is no 16 bit float format supported by any existing hardware,
Sure, they're called GPU's :) At CGShaders.org, there's even a raytraced
pixel shader, though you have to compile it separately for each scene.
...Chambers
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Christopher James Huff wrote:
>
>>Would it be possible/easy to compile a version of Povray which uses 16
>>bit floats?
>
>
> Why would you want a version that used 16 bit floats?
>
Embedded Povray ?
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
In article <3f20e59f@news.povray.org>, Th Boudet <th.### [at] freefr>
wrote:
> Christopher James Huff wrote:
> > Why would you want a version that used 16 bit floats?
> Embedded Povray ?
And why would you want that?
Anyway, it just won't work. You could write a simple custom raytracer if
you pay very careful attention to precision, but POV has enough
precision problems already.
--
Christopher James Huff <cja### [at] earthlinknet>
http://home.earthlink.net/~cjameshuff/
POV-Ray TAG: chr### [at] tagpovrayorg
http://tag.povray.org/
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
