Most modern PCs have a very powerful 3d video card with a GPU that is
hundreds if not thousands of times faster than the CPU for certain
calculations. I was thinking if some of the POV computations could be
offloaded to the GPU, we could speed up renders by a very large factor. The
big question is one of which routines thouse would be and which chunks of
the GPU could be utilized. Certainly matrix operations could be offloaded at
the very least, although I don't know to what extent that would improve
performance. It's just a shame that there is this uber-powerful floating
point processor sitting in most PCs that can't be used to help things move
along. What do others think?

Post a reply to this message

In article <3ee74af9$2@news.povray.org> , "Michael Goldshteyn" 
<mgo### [at] n-o-s-p-a-m-earthlinknet> wrote:

> Most modern PCs have a very powerful 3d video card with a GPU that is
> hundreds if not thousands of times faster than the CPU for certain
> calculations. I was thinking if some of the POV computations could be
> offloaded to the GPU, we could speed up renders by a very large factor. The
> big question is one of which routines thouse would be and which chunks of
> the GPU could be utilized. Certainly matrix operations could be offloaded at
> the very least, although I don't know to what extent that would improve
> performance. It's just a shame that there is this uber-powerful floating
> point processor sitting in most PCs that can't be used to help things move
> along. What do others think?

You need double precision floating-point processing.  Sorry!

    Thorsten

Post a reply to this message

"Thorsten Froehlich" <tho### [at] trfde> wrote in message
news:3ee74bb0$1@news.povray.org...
> In article <3ee74af9$2@news.povray.org> , "Michael Goldshteyn"
> <mgo### [at] n-o-s-p-a-m-earthlinknet> wrote:
>
> > Most modern PCs have a very powerful 3d video card with a GPU that is
> > hundreds if not thousands of times faster than the CPU for certain
> > calculations. I was thinking if some of the POV computations could be
> > offloaded to the GPU, we could speed up renders by a very large factor.
The
> > big question is one of which routines thouse would be and which chunks
of
> > the GPU could be utilized. Certainly matrix operations could be
offloaded at
> > the very least, although I don't know to what extent that would improve
> > performance. It's just a shame that there is this uber-powerful floating
> > point processor sitting in most PCs that can't be used to help things
move
> > along. What do others think?
>
> You need double precision floating-point processing.  Sorry!
>
>     Thorsten

Sorry, I accidentally e-mailed my reply, instead of posting it.

In any case, a quote from NVIDIA's web site:

And, powered by the second-generation NVIDIA CineFXT 2.0 engine with the
industry's only true 128-bit precision processing--the GeForce FX 5900 GPUs
take cinematic-quality special effects to new levels while providing the
industry's most compatible and reliable gaming platform.

Now, double-precision uses 64-bit floating point, so I think this is more
than enough.

Post a reply to this message

Thorsten Froehlich wrote:
> You need double precision floating-point processing.  Sorry!

I wanted answer almost identicall but I'm completly dumb in that area so did
quick search in google. It answered there are GPUs with 128-bit
floating-point precision.

ABX

Post a reply to this message

After doing more reading, the 128 bit precision may refer to 32-bit floating
point times four. So, in fact, the GPU may only be single-precision per
color.

Post a reply to this message

In article <web.3ee74ea3d3e4b3b1465fc1770@news.povray.org> , 
"Włodzimierz ABX Skiba" <abx### [at] abxartpl> wrote:

> I wanted answer almost identicall but I'm completly dumb in that area so did
> quick search in google. It answered there are GPUs with 128-bit
> floating-point precision.

No, you simply fall for a good trap their marketing people have set up.  The
GPUs are SIMD processors and the units work with four single precision
floating-point numbers*.  Hence, for a marketing person it works with 128
bits precision using floating-point numbers <sigh>

    Thorsten

* The smallest document on the Nvidia site that details the marketing math
is <http://developer.nvidia.com/docs/IO/3121/ATT/CineFX-TechBrief_v02.pdf>

____________________________________________________
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

Thorsten Froehlich wrote:
> You need double precision floating-point processing.  Sorry!

But do you *really* need it for everything

-- 
Rick

Kitty5 NewMedia http://Kitty5.com
POV-Ray News & Resources http://Povray.co.uk
TEL : +44 (01270) 501101 - FAX : +44 (01270) 251105 - ICQ : 15776037

PGP Public Key
http://pgpkeys.mit.edu:11371/pks/lookup?op=get&search=0x231E1CEA

Post a reply to this message

On Wed, 11 Jun 2003 19:31:17 +0100, "Rick [Kitty5]" <spa### [at] kitty5com> wrote:
> > You need double precision floating-point processing.  Sorry!
>
> But do you *really* need it for everything

Single precision floats are used for colors in POV sources.

ABX

Post a reply to this message

In article <3ee77604$1@news.povray.org> , "Rick [Kitty5]" 
<spa### [at] kitty5com> wrote:

>> You need double precision floating-point processing.  Sorry!
>
> But do you *really* need it for everything

You need it for intersection tests.  You cannot store enough data in 32 bits
to perform adequate intersection tests without limiting the geometry.

You seem to misunderstand what precision anything stored in 32 bits can
possibly provide.  From a more conventional perspective, I would suggest to
get a introduction to combinatorics book and read the chapter on the
pigeonhole principle.

    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 <3ee74af9$2@news.povray.org>,
 "Michael Goldshteyn" <mgo### [at] n-o-s-p-a-m-earthlinknet> wrote:

> Most modern PCs have a very powerful 3d video card with a GPU that is
> hundreds if not thousands of times faster than the CPU for certain
> calculations. I was thinking if some of the POV computations could be
> offloaded to the GPU, we could speed up renders by a very large factor. The
> big question is one of which routines thouse would be and which chunks of
> the GPU could be utilized. Certainly matrix operations could be offloaded at
> the very least, although I don't know to what extent that would improve
> performance. It's just a shame that there is this uber-powerful floating
> point processor sitting in most PCs that can't be used to help things move
> along. What do others think?

As others have mentioned, because of precision issues, this isn't useful 
for geometry calculations. But more importantly, this isn't like a 
floating point coprocessor, it is a video card, and is designed to 
efficiently display things. Even if the card can do the operations 
faster, it will take far more time to send the data to the card to be 
processed and then retrieve it than it would to process the data on the 
CPU. There are some things it could accelerate, but typically not what 
POV needs.
With more programmable cards becoming available, you can do more 
general-purpose things with the card...it may be possible to process 
data in large enough chunks that it is worth it to use the card, maybe 
even doing all the raytracing on the card. However, you'd most likely 
have to design the raytracer around this, I don't see anyone converting 
POV to do it. And then you have the hardware dependencies, etc...

-- 
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