 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Having taken a diff,
http://www.ladylittlefox.com/10_abs_32-64.jpg
, I can
see several artifacts that are easily attributable to random numbers,
specifically the aliased edges of approximately everything. However, there
are huge differences at:
1) the inner edge of the annular caustic around pixel {490,490}
(unexplained sharp edge in "32 bits")
2) the inner face of the left-rear retaining wall
(much brighter in "32 bits")
3) the floor under the cage in the lower-right
(too hard to tell what's happening by eye, so probably not important)
The above all appear to be in shadow volumes, but the next one does not.
4) a series of (differentially) red and green spots on the illuminated side
of the "rocky" cylinder.
(easy to see in the "32 bits" image when you know where to look)
and
5) The set of jitter errors in the marbling on the lower-left wall appear to
be organized on a perspective grid. This may just be an illusion.
6) The translucent sphere in the top-right also has a high energy feature.
It's not clear if the shadow feature is jitter on both edges that are so
close together that it't "all jitter". However, the light on the bezel
behind it is much dimmer in the "32 bits" image.
-- Eric Towers
fuz### [at] bigfoot com
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Seeing the actual images would be interesting as well.
I mean, are the differences clear bugs, or do both images just look ok,
only slightly different?
--
#macro N(D)#if(D>99)cylinder{M()#local D=div(D,104);M().5,2pigment{rgb M()}}
N(D)#end#end#macro M()<mod(D,13)-6mod(div(D,13)8)-3,10>#end blob{
N(11117333955)N(4254934330)N(3900569407)N(7382340)N(3358)N(970)}// - Warp -
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Warp nous apporta ses lumieres ainsi en ce 10/06/2004 15:24... :
> Seeing the actual images would be interesting as well.
> I mean, are the differences clear bugs, or do both images just look ok,
>only slightly different?
>
>
>
Here are the urls for the images:
32 bits
http://www.linuxhardware.org/images/articles/opteron-121703/bench-i386-full.jpg
64 bits
http://www.linuxhardware.org/images/articles/opteron-121703/bench-x86-64-full.jpg
Alain
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
There's certainly a visible difference in the photon mapping.
In the 64-bit version the photon mapping seems to be smoother.
There might be some code in the photon mapping routines which
behaves differently when compiled in 64-bit mode.
Btw, which version of POV-Ray did you use to render these images?
And what about other system specs?
It will be interesting to see if there's any difference with POV-Ray 3.6
(once the source code is published) since some modifications were made
to the photon mapping routines...
--
#macro N(D)#if(D>99)cylinder{M()#local D=div(D,104);M().5,2pigment{rgb M()}}
N(D)#end#end#macro M()<mod(D,13)-6mod(div(D,13)8)-3,10>#end blob{
N(11117333955)N(4254934330)N(3900569407)N(7382340)N(3358)N(970)}// - Warp -
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Warp nous apporta ses lumieres ainsi en ce 11/06/2004 09:03... :
> There's certainly a visible difference in the photon mapping.
>In the 64-bit version the photon mapping seems to be smoother.
>There might be some code in the photon mapping routines which
>behaves differently when compiled in 64-bit mode.
>
> Btw, which version of POV-Ray did you use to render these images?
>And what about other system specs?
>
> It will be interesting to see if there's any difference with POV-Ray 3.6
>(once the source code is published) since some modifications were made
>to the photon mapping routines...
>
>
>
I have nothing to do with those images, I only posted the links. I don't
have an opteron, just an oldish AMD 1400...
Alain
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
