Has any P4 owner noticed speedups with POVRay 3.5 RC5 ? I noticed in the
changes log that they started to add optimizations, but it doesnt say much
else. I can't test this on my Athlon TBird :)

George Pantazopoulos

Post a reply to this message

George Pantazopoulos <the### [at] attbicom> wrote:
> Has any P4 owner noticed speedups with POVRay 3.5 RC5 ? I noticed in the
> changes log that they started to add optimizations, but it doesnt say much
> else. I can't test this on my Athlon TBird :)

  It is quite common that code that has been optimized for P4 also runs
faster in an Athlon than the unoptimized code, so some improvement should
probably be seen.

-- 
#macro M(A,N,D,L)plane{-z,-9pigment{mandel L*9translate N color_map{[0rgb x]
[1rgb 9]}scale<D,D*3D>*1e3}rotate y*A*8}#end M(-3<1.206434.28623>70,7)M(
-1<.7438.1795>1,20)M(1<.77595.13699>30,20)M(3<.75923.07145>80,99)// - Warp -

Post a reply to this message

On 20 May 2002 11:00:09 -0400, Warp <war### [at] tagpovrayorg> wrote:

>George Pantazopoulos <the### [at] attbicom> wrote:
>> Has any P4 owner noticed speedups with POVRay 3.5 RC5 ? I noticed in the
>> changes log that they started to add optimizations, but it doesnt say much
>> else. I can't test this on my Athlon TBird :)
>
>  It is quite common that code that has been optimized for P4 also runs
>faster in an Athlon than the unoptimized code, so some improvement should
>probably be seen.

I know that it's not totally fair to compare RC4 with the new RC5,
with respect to the new optimizations, but on an AthlonXP 1600+
system, I have a scene file that renders one second slower with the
new compile.

RC4 = 56 seconds
RC5 = 57 seconds

This is only a single file, selected at random, and far from a
scientific evaluation. However, I mention it to point out that the
potential gains from the new optimizations don't always happen for all
scene files, at least not on an AthlonXP system. I'm not alarmed by
this, especially since the new performance gains were only reported to
be moderate anyway, even when running under an Intel processor, as
targeted by these optimizations. 

I'll run a few more comparisons, and if I find anything remarkable
happening, I'll be sure to post something about it. 

I wonder which group that should be posted in, when and if I feel the
need to do so?



Later,
Glen

7no### [at] ezwvcom     (Remove the numeral "7")

Post a reply to this message

In article <KfnpPGvd4oUpnzF6gIpfPaCyOZVg@4ax.com> , Glen Berry 
<7no### [at] ezwvcom>  wrote:

> I know that it's not totally fair to compare RC4 with the new RC5,
> with respect to the new optimizations, but on an AthlonXP 1600+
> system, I have a scene file that renders one second slower with the
> new compile.
>
> RC4 = 56 seconds
> RC5 = 57 seconds

You know, on a multitasking system one second is like nothing.  A measuring
error, not more.

To trace the effect of the speedup you need a scene which uses a lot of noise,
that is for the most part certain textures.  To try benchmark.pov for example
(with a small resolution) will show some differences.

    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