POV-Ray : Newsgroups : povray.newusers : interpreting pov statistics Server Time
5 Nov 2024 11:19:54 EST (-0500)
  interpreting pov statistics (Message 1 to 3 of 3)  
From: Bill Hails
Subject: interpreting pov statistics
Date: 4 Nov 2003 03:35:17
Message: <3fa764c5@news.povray.org>
Hi,
I just waited 18 hours for a 320 x 240 snapshot to render, and that doesn't 
bode well for any larger sizes. Mostly it seems that an area light is what 
slowed it down so much, but I'd appreciate it if some kind soul could 
explain to me how to interpret the statistics that pov prints at the end of 
a run, so I can understand how I might speed things up. I'm using 
radiosity, and a saved photon map from a previous run.

Here's what I got:


Statistics for Orrery.pov, Resolution 320 x 240
----------------------------------------------------------------------------
Pixels:           77790   Samples:          299294   Smpls/Pxl: 3.85
Rays:           1916709   Saved:              4059   Max Level: 24/255
----------------------------------------------------------------------------
Ray->Shape Intersection          Tests       Succeeded  Percentage
----------------------------------------------------------------------------
Blob                          16563127         1082976      6.54
Blob Component               414964724        19817115      4.78
Blob Bound                  1177489001       498256364     42.32
Box                           79375558        30035034     37.84
Cone/Cylinder                120724632        76992925     63.78
CSG Intersection              96267462        32469328     33.73
CSG Merge                     11742121          480810      4.09
CSG Union                     26082309         2089961      8.01
Fractal                      315767336        28621239      9.06
Height Field                  59749281        29959847     50.14
Height Field Box              59749281        59635002     99.81
Height Field Triangle         92811896        29976415     32.30
Height Field Block            59759349        39016074     65.29
Height Field Cell             47903328        46412956     96.89
Isosurface                   311928365       253040365     81.12
Isosurface Container         315506286       311930882     98.87
Isosurface Cache                375029           22959      6.12
Lathe                          2124622          136523      6.43
Lathe Bound                    2124622          237868     11.20
Plane                        823273396       383136223     46.54
Prism                         13479082         2201607     16.33
Prism Bound                   54376708        33993342     62.51
Sphere                       326642099       180434182     55.24
Sphere Sweep                 138385042         3270646      2.36
Superellipsoid                66983595         9375622     14.00
Torus                         10579384         1633723     15.44
Torus Bound                   10579384         1969980     18.62
True Type Font                 1338511           80381      6.01
Bounding Object                 862690          144690     16.77
Bounding Box                9337975026      3204610846     34.32
Vista Buffer                  27184827        18665289     68.66
----------------------------------------------------------------------------
Isosurface roots:        311927311
Function VM calls:     13539920339
----------------------------------------------------------------------------
Roots tested:            364414141   eliminated:              593946
Calls to Noise:        13563987641   Calls to DNoise:       61099180
----------------------------------------------------------------------------
Media Intervals:          24149072   Media Samples:         34447804 (1.43)
Shadow Ray Tests:        240890798   Succeeded:             19112185
Reflected Rays:             168618
Refracted Rays:              36645
Transmitted Rays:           862140
Priority queue insert:          318263
Priority queue remove:           90183
Gather function called:        1642579
Gather radius expanded:         755703
----------------------------------------------------------------------------
Radiosity samples calculated:      18092 (1.27 percent)
Radiosity samples reused:        1408326
----------------------------------------------------------------------------
Smallest Alloc:                  9 bytes   Largest:         16000008
Peak memory used:        150850311 bytes
----------------------------------------------------------------------------
Time For Parse:    0 hours  0 minutes  10.0 seconds (10 seconds)
Time For Photon:   0 hours  0 minutes   3.0 seconds (3 seconds)
Time For Trace:   18 hours  2 minutes   2.0 seconds (64922 seconds)
    Total Time:   18 hours  2 minutes  15.0 seconds (64935 seconds)

Any help gratefully accepted.

-- 
Bill Hails


Post a reply to this message

From: Warp
Subject: Re: interpreting pov statistics
Date: 4 Nov 2003 07:57:43
Message: <3fa7a247@news.povray.org>
Bill Hails <bil### [at] europeyahoo-inccom> wrote:
> Blob                          16563127         1082976      6.54
> Blob Component               414964724        19817115      4.78
> Blob Bound                  1177489001       498256364     42.32

  Try manually bounding your blob with a tighter sphere or square.
The hitrate above is very low and that's because POV-Ray usually
creates very big boinding boxes for blobs (it's difficult to get
an optimal one, even though I don't know if it's impossible).

  You can find an optimal box or sphere by rendering a semi-transparent
one around your blob and fine-tuning its location and dimensions until
it tightly fits the blob. Then just add a bounding box or sphere to
the blob with the same parameters.

> Fractal                      315767336        28621239      9.06
> Lathe                          2124622          136523      6.43
> Sphere Sweep                 138385042         3270646      2.36

  The same advice goes for these as well.


  You can often fine-tune your area light to render slightly faster
by modifying its parameters (mainly the 'adaptive' value, but sometimes
even lowering the density of the pointlights). If the area light
is lighting only a limited section of your scene, try seeing if you
can make it a spotlight.

-- 
plane{-x+y,-1pigment{bozo color_map{[0rgb x][1rgb x+y]}turbulence 1}}
sphere{0,2pigment{rgbt 1}interior{media{emission 1density{spherical
density_map{[0rgb 0][.5rgb<1,.5>][1rgb 1]}turbulence.9}}}scale
<1,1,3>hollow}text{ttf"timrom""Warp".1,0translate<-1,-.1,2>}//  - Warp -


Post a reply to this message

From: Bill Hails
Subject: Re: interpreting pov statistics
Date: 4 Nov 2003 15:26:36
Message: <3fa80b7c@news.povray.org>
Warp wrote:

> Bill Hails <bil### [at] europeyahoo-inccom> wrote:
>> Blob                          16563127         1082976      6.54
>> Blob Component               414964724        19817115      4.78
>> Blob Bound                  1177489001       498256364     42.32
> 
>   Try manually bounding your blob with a tighter sphere or square.
> The hitrate above is very low and that's because POV-Ray usually
> creates very big boinding boxes for blobs (it's difficult to get
> an optimal one, even though I don't know if it's impossible).

Ah, I see it.

>   You can find an optimal box or sphere by rendering a semi-transparent
> one around your blob and fine-tuning its location and dimensions until
> it tightly fits the blob. Then just add a bounding box or sphere to
> the blob with the same parameters.
> 
>> Fractal                      315767336        28621239      9.06
>> Lathe                          2124622          136523      6.43
>> Sphere Sweep                 138385042         3270646      2.36
> 
>   The same advice goes for these as well.

Yes, thanks.

> 
>   You can often fine-tune your area light to render slightly faster
> by modifying its parameters (mainly the 'adaptive' value, but sometimes
> even lowering the density of the pointlights). If the area light
> is lighting only a limited section of your scene, try seeing if you
> can make it a spotlight.
> 

Spotlight is not an option, I'm afraid, but I'll take a look at 'adaptive'.

Thanks very much for your help.
-- 
Bill Hails


Post a reply to this message

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