Tearing down the scene to investigate render time problems.

Trace time on this pass was something over 14h, stats are attached below.

I would have expected it to render faster with just one thread per core, 
but that doesn't seem to be the case.  This render used 48 threads.

I'm using Radiosity_IndoorHQ from rad_def.inc with one small(ish) 
alteration.  Since there aren't any light sources, I bumped the count up 
to 160k and set a default importance of .001 with the emitting objects 
having an importance of 1.

one problem area seems to be the tiles.  I only have one patch of tiles 
in this render (the smallest one).

I said in the other thread that the tiles are modeled as a single prism, 
but I was wrong (obviously, since there are two different colors) 
they're actually a merge of prisms, each one is individually textured.

I'd consider breaking them into two sets and texturing each set as a 
whole, except that it might look odd if the texture for each tile 
weren't different.

Regards,
A.D.B.

Post a reply to this message

Attachments:
Download 'csgtest01.png' (192 KB) Download 'stats.txt' (9 KB)

Preview of image 'csgtest01.png'

On 09/04/2015 12:14 AM, Anthony D. Baye wrote:
> I said in the other thread that the tiles are modeled as a single prism,
> but I was wrong (obviously, since there are two different colors)
> they're actually a merge of prisms, each one is individually textured.

Is there a reason you are using a merge rather than a union?  It is hard 
for me to see what is going on, due to the low lighting.

http://wiki.povray.org/content/Knowledgebase:Miscellaneous#Topic_3

Post a reply to this message

Cousin Ricky <ric### [at] yahoocom> wrote:
> On 09/04/2015 12:14 AM, Anthony D. Baye wrote:
> > I said in the other thread that the tiles are modeled as a single prism,
> > but I was wrong (obviously, since there are two different colors)
> > they're actually a merge of prisms, each one is individually textured.
>
> Is there a reason you are using a merge rather than a union?  It is hard
> for me to see what is going on, due to the low lighting.
>
> http://wiki.povray.org/content/Knowledgebase:Miscellaneous#Topic_3

I believe it was faster to use a merge when I was first testing the tile macro.

Regards,
A.D.B.

Post a reply to this message

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Le 04/09/2015 21:00, Anthony D. Baye a écrit :
> Cousin Ricky <ric### [at] yahoocom> wrote:
>> On 09/04/2015 12:14 AM, Anthony D. Baye wrote:
>>> I said in the other thread that the tiles are modeled as a
>>> single prism, but I was wrong (obviously, since there are two
>>> different colors) they're actually a merge of prisms, each one
>>> is individually textured.
>> 
>> Is there a reason you are using a merge rather than a union?  It
>> is hard for me to see what is going on, due to the low lighting.
>> 
>> http://wiki.povray.org/content/Knowledgebase:Miscellaneous#Topic_3
>
>> 
> I believe it was faster to use a merge when I was first testing the
> tile macro.

not really. Merge need to compute all the intersections of a ray with
the object, then remove the ones that are inside the object.

which means an additional inside test per intersection point, IIRC. If
the texture is opaque, union is enough (as it does not remove the
internal points)

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iJwEAQEIAAYFAlXqj6gACgkQhKAm8mTpkW1CywQA3YklEgHJChlxBWKHWu4W8VpV
DMB/ZxstClAd9LtkoWRUiTm51bLSk8vx/LRE7QYWc+PR3LS1zTdlyM1j7XUbiGVW
8mP4Pv3wZnQkJOtfXp+lbD09HC8mY+k7NGJr42UhfrWyor31Mma+XzxKsEemkx8X
DbsRwGNhFkFva+masWA=
=3Ds9
-----END PGP SIGNATURE-----

Post a reply to this message