In article <3e655383$1@news.povray.org> , "Ian J. Burgmyer" 
<the### [at] maccom> wrote:

> Right...that's why it's reproducable on my system (and yes, MacMegaPOV works
> fine).  I haven't tried the 3.5.1 preview yet, so I can't say anything about
> that.
>
> Followups set to povray.macintosh.

So you say with preview off and preemptive threading on, scenes that are not
subject to the speed improvements in MegaPOV, which are for obvious versions
not included in the official POV-Ray 3.5 as they didn't exist back then,
there is a speed difference for a scene that takes more than an 10 minutes
to render? (anything shorter wouldn't be a useful measurement)

    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

Thorsten Froehlich's furious key-hammering produced this:
> So you say with preview off and preemptive threading on, scenes that are not
> subject to the speed improvements in MegaPOV, which are for obvious versions
> not included in the official POV-Ray 3.5 as they didn't exist back then,
> there is a speed difference for a scene that takes more than an 10 minutes
> to render? (anything shorter wouldn't be a useful measurement)

The problem isn't that rendering is slow, it's that POV-Ray is only eating up a
fraction of CPU time (I did a check with no other active applications using top
and it only seems to utilize a maximum of 25%).

OTOH, I haven't tried preemptive threading yet.

-- 
/*^*/light_source{100*<-5,2,-5>2}#macro I(i,n)#while(strlen(i)>=n)#local A=asc(
substr(i,n,1));#local a=asc(substr(i,n+1,1));cylinder{<div(A,8)-12,mod(A,8)-4,4
><div(a,8)-12,mod(a,8)-4,4>,0.1pigment{rgb z}}#local n=n+2;#end#end I("ScUe[]"1
/*<*/)I("mkmtlttk"1)//@_$#!,:<"Thhis polysig brought to you by Ian Burgmyer :)"

Post a reply to this message

In article <3e66b21f@news.povray.org> , "Ian J. Burgmyer" 
<the### [at] maccom> wrote:

> The problem isn't that rendering is slow, it's that POV-Ray is only eating up
a
> fraction of CPU time (I did a check with no other active applications using
top
> and it only seems to utilize a maximum of 25%).

Well, why do you worry about some incorrect display, that isn't even meant
for what you are using it (not suitable at all), tell you?  The only thing
of interest is the result, which is the render and parsing time, which does
not depend on what some stupid display tells you!

Hmm, maybe I should divide render times displayed by two and people will
think POV-Ray got twice as fast without ever checking? ;-) ;-) ;-)

    Thorsten

____________________________________________________
Thorsten Froehlich
e-mail: mac### [at] povrayorg

I am a member of the POV-Ray Team.
Visit POV-Ray on the web: http://mac.povray.org

Post a reply to this message

Thorsten Froehlich's furious key-hammering produced this:
>> The problem isn't that rendering is slow, it's that POV-Ray is only eating up
>> a
>> fraction of CPU time (I did a check with no other active applications using
>> top
>> and it only seems to utilize a maximum of 25%).
> 
> Well, why do you worry about some incorrect display, that isn't even meant
> for what you are using it (not suitable at all), tell you?  The only thing
> of interest is the result, which is the render and parsing time, which does
> not depend on what some stupid display tells you!

That's the thing -- because of it only using a fraction of my CPU time it takes
the system a ridiculous amount of time to render a very simple scene.  I don't
remember specifics, but my Pentium III 750 system can render very basic lit
sphere scene (source below) in a second.  OTOH, my G4-733MHz (which actually
renders faster than the x86 when both platforms are using MegaPOV) took over
15 seconds to render it.

This is the scene I used.
light_source{<15,8,-25>rgb 1}sphere{z*5,1pigment{rgb 1}}

As I said, I haven't tried 3.5.1 (though I'm sure that one will work much
better), only 3.5.0 and MacMegaPOV.  I'll have to remind myself to give that a
try over the weekend...

-- 
/*^*/light_source{100*<-5,2,-5>2}#macro I(i,n)#while(strlen(i)>=n)#local A=asc(
substr(i,n,1));#local a=asc(substr(i,n+1,1));cylinder{<div(A,8)-12,mod(A,8)-4,4
><div(a,8)-12,mod(a,8)-4,4>,0.1pigment{rgb z}}#local n=n+2;#end#end I("ScUe[]"1
/*<*/)I("mkmtlttk"1)//@_$#!,:<"Thhis polysig brought to you by Ian Burgmyer :)"

Post a reply to this message

> That's the thing -- because of it only using a fraction of my CPU time it
takes
> the system a ridiculous amount of time to render a very simple scene.  I
don't
> remember specifics, but my Pentium III 750 system can render very basic
lit
> sphere scene (source below) in a second.  OTOH, my G4-733MHz (which
actually
> renders faster than the x86 when both platforms are using MegaPOV) took
over
> 15 seconds to render it.
>
> This is the scene I used.
> light_source{<15,8,-25>rgb 1}sphere{z*5,1pigment{rgb 1}}

With preview on and with preemptive threading (call background rendering in
3.51) off, right?

Note that the time it takes to draw the preview remains constant no matter
what scene complexity.  Hence, it will still be 15 seconds for a 10 minute
scene, at which point 15 seconds added time for preview don't matter.  That
is why one does not benchmark with scenes that take one second to render,
because that ain't a benchmark, but a bananamark ;-)

    Thorsten

Post a reply to this message

Thorsten Froehlich's furious key-hammering produced this:
>> sphere scene (source below) in a second.  OTOH, my G4-733MHz (which
>> actually renders faster than the x86 when both platforms are using
>> MegaPOV) took over 15 seconds to render it.
>>
>> This is the scene I used.
>> light_source{<15,8,-25>rgb 1}sphere{z*5,1pigment{rgb 1}}
> 
> With preview on and with preemptive threading (call background rendering in
> 3.51) off, right?

Yes -- default options, basically.  I wasn't feeling too adventurous when I did
that quick test. :)

> Note that the time it takes to draw the preview remains constant no matter
> what scene complexity.  Hence, it will still be 15 seconds for a 10 minute
> scene, at which point 15 seconds added time for preview don't matter.

Oh, alright.  So you're saying that with preview off that scene would have
probably rendered as fast as my Wintel rendered it?  Sounds plausible to me.

Heh, I still have to try 3.5.1...

> That
> is why one does not benchmark with scenes that take one second to render,
> because that ain't a benchmark, but a bananamark ;-)

Bananamark...hehe, I like that. ;)

I'll have to run benchmark.pov on both of my systems to see which one is truly
a better POV-platform (with preview off, of course!).  I honestly never thought
that preview being on made that much of a difference!

-- 
/*^*/light_source{100*<-5,2,-5>2}#macro I(i,n)#while(strlen(i)>=n)#local A=asc(
substr(i,n,1));#local a=asc(substr(i,n+1,1));cylinder{<div(A,8)-12,mod(A,8)-4,4
><div(a,8)-12,mod(a,8)-4,4>,0.1pigment{rgb z}}#local n=n+2;#end#end I("ScUe[]"1
/*<*/)I("mkmtlttk"1)//@_$#!,:<"Thhis polysig brought to you by Ian Burgmyer :)"

Post a reply to this message

In article <3e6af26e$1@news.povray.org> , "Ian J. Burgmyer" 
<the### [at] maccom> wrote:

> I'll have to run benchmark.pov on both of my systems to see which one is truly
> a better POV-platform (with preview off, of course!).  I honestly never
thought
> that preview being on made that much of a difference!

Now, this has been discussed regarding Mac OS X focred double-buffered
windows before in this group more; than once.

    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

Thorsten Froehlich's furious key-hammering produced this:
>> I'll have to run benchmark.pov on both of my systems to see which one is truly
>> a better POV-platform (with preview off, of course!).  I honestly never
>> thought that preview being on made that much of a difference!
> 
> Now, this has been discussed regarding Mac OS X focred double-buffered
> windows

Forced double-buffering??  Say no more...

> before in this group more; than once.

I only started reading this group recently and my newsreader only shows new
posts.

-- 
/*^*/light_source{100*<-5,2,-5>2}#macro I(i,n)#while(strlen(i)>=n)#local A=asc(
substr(i,n,1));#local a=asc(substr(i,n+1,1));cylinder{<div(A,8)-12,mod(A,8)-4,4
><div(a,8)-12,mod(a,8)-4,4>,0.1pigment{rgb z}}#local n=n+2;#end#end I("ScUe[]"1
/*<*/)I("mkmtlttk"1)//@_$#!,:<"Thhis polysig brought to you by Ian Burgmyer :)"

Post a reply to this message