 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hi all,
If you look at the end of a review of Opterons at
http://www.linuxhardware.org/Features/03/12/17/189239.shtml, you can see
benchmarks done with Pov-Ray compiled for x86-64 and for regular x86,
using the same scene description. The images are non-identical, and I'd
like to know why this is; is it just because all calculations are done
with 64 bit precision, as the article says, or are there other
differences in Pov-Ray's behaviour on x86-64 systems?
Thanks,
Nathan Cook
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
In article <40b87180$1@news.povray.org> , use### [at] domain invalid wrote:
> If you look at the end of a review of Opterons at
> http://www.linuxhardware.org/Features/03/12/17/189239.shtml, you can see
> benchmarks done with Pov-Ray compiled for x86-64 and for regular x86,
> using the same scene description. The images are non-identical, and I'd
> like to know why this is; is it just because all calculations are done
> with 64 bit precision, as the article says, or are there other
> differences in Pov-Ray's behaviour on x86-64 systems?
Their observation that POV-Ray yields slightly different results is correct.
Actually, this is the case whenever you compare renders on any two different
processors. It is also true that comparing a 32 bit and a 64 bit compile on
POV-Ray on an AMD x86 system running in 64 bit mode yields differences.
However, the statement
"It seems that POV-Ray was not written with fixed precision, so the image on
the right has greater precision in its calculations than the left one. This
is apparent most in the shadowing. If you zoom in on shadowed areas, you'll
notice a smoother gradient in the colors. From these results we see that not
only do you get greater precision for free, it's actually "cheaper" and
completes faster than the 32-bit version."
is pure nonsense. No fact at all, just random speculation by someone who
obviously has no understanding of the benchmark he ran. The "64 bit" refer
to the integer unit and the memory addressing capability of a processor,
they have absolutely nothing to do with the floating-point capabilities of a
processor.
To actually explain the differences generated on the same processor when
running in 64 bit vs. 32 bit mode, one needs (unavailable!) details about
the floating-point precision provided by the logic on the processor itself
as well as possible differences of code generated by the compiler used when
targeting a different instruction (sub)set. For certain operations like
division and square-root computation of floating-point numbers, only a
certain minimum precision is required. This is in particular relevant when
performing many computation in series. POV-Ray does not perform rounding of
those results, and as such always uses the maximum floating-point precision
provided by the hardware.
Nevertheless, a floating-point number still has the exactly same size and
bit representation regardless of the processor being "64 bit" or "32 bit"
because as said above already, the "64 bit" only refers to the integer size
supported. In order to understand why floating-point numbers do only have a
certain precision and the order of computations can change that noticeably,
in-depth understanding of the math behind them is needed. So I am not going
to explain this here.
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
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Actually, I think I did not make clear one other important thing:
The result when running in the AMD 64 bit mode is just *different*, not
better or worse than the AMD 32 bit mode result. As such, there is in
particular *no* higher quality as the result of AMD 64 bit mode!
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
|
|
| |
| |
|
|
|
|
| |
| |
|
|
use### [at] domaininvalid wrote:
> Hi all,
>
> If you look at the end of a review of Opterons at
> http://www.linuxhardware.org/Features/03/12/17/189239.shtml, you can see
> benchmarks done with Pov-Ray compiled for x86-64 and for regular x86,
> using the same scene description. The images are non-identical, and I'd
> like to know why this is; is it just because all calculations are done
> with 64 bit precision, as the article says, or are there other
> differences in Pov-Ray's behaviour on x86-64 systems?
If that are really true results of the official benchmark (i.e. not made
with other render options) i find the results quite bad for such
expensive systems. Of course a hardware like that is quite
'underemployed' with rendering the POV-Ray benchmark but the Xeon for
example seems not much faster than my Athlon XP 2600 here and you
probably could get a dozen of them for the same price as the Xeon...
The 22:20 for the Opteron is of course not that bad but you can achieve
similar results on much cheaper hardware, i.e. Athlon 64:
http://www.linuxhardware.org/article.pl?sid=04/03/18/2143203
Christoph
--
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 01 May. 2004 _____./\/^>_*_<^\/\.______
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Thorsten Froehlich <tho### [at] trfde> wrote:
> Their observation that POV-Ray yields slightly different results is correct.
> Actually, this is the case whenever you compare renders on any two different
> processors. It is also true that comparing a 32 bit and a 64 bit compile on
> POV-Ray on an AMD x86 system running in 64 bit mode yields differences.
But what is the source of these differences?
They definetely don't look like caused by different rounding, or by a different
random number generator or anything like that.
Actually it looks more as if different setting for photon mapping where used.
If the settings where identical I suspect there might be a bug in the photon
mapping code.
Lutz-Peter
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
In article <40bb89f8@news.povray.org> , Lutz-Peter Hooge <lpv### [at] gmxde>
wrote:
> But what is the source of these differences?
Some POV-Ray features use pseudo-random numbers. The sequence of these
numbers depends on the order of certain tracing calls. If there is a small
computational difference in just one pixel POV-Ray may need to trace
additional rays (i.e. for anti-aliasing, but not limited to it*), you get a
butterfly effect when computing all consecutive pixels...
Thorsten
* Chris Cason recently found that using jitter is very sensitive to the
butterfly effect.
____________________________________________________
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 <tho### [at] trfde> wrote:
> Some POV-Ray features use pseudo-random numbers. The sequence of these
> numbers depends on the order of certain tracing calls. If there is a small
IMO this would result mainly in differences at sharp edges etc.
The main differences in the shown images are of a different type,
it is most visible in the bright caustics in the middle ofthe image
http://www.linuxhardware.org/images/articles/opteron-121703/bench-x86-64-full.jpg
http://www.linuxhardware.org/images/articles/opteron-121703/bench-i386-full.jpg
Lutz-Peter
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Lutz-Peter Hooge" <lpv### [at] gmxde> schrieb im Newsbeitrag
news:40bbef01$1@news.povray.org...
> Thorsten Froehlich <tho### [at] trfde> wrote:
>
> > Some POV-Ray features use pseudo-random numbers. The sequence of these
> > numbers depends on the order of certain tracing calls. If there is a
small
>
> IMO this would result mainly in differences at sharp edges etc.
> The main differences in the shown images are of a different type,
> it is most visible in the bright caustics in the middle ofthe image
>
>
http://www.linuxhardware.org/images/articles/opteron-121703/bench-x86-64-full.jpg
>
http://www.linuxhardware.org/images/articles/opteron-121703/bench-i386-full.jpg
Well, what I said also goes for photons. Plus, there is no way to rule you
they didn't make some mistakes when running the benchmark. All I can really
say is that their reasoning that using a 64 bit compile creates "better" is
not based on any facts. It is certainly possible to make good guesses what
else might be wrong with their compile, but those don't really lead
anywhere, I am afraid.
Thorsten
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
FYI:
http://www.povray.org/temp/benchmark32.png
http://www.povray.org/temp/benchmark64.png
http://www.povray.org/temp/32-64-diff.png
rendered with 32 and 64-bit versions of POV-Ray v3.6 on an Athlon 64 FX-51
running XP/64. please don't ask about benchmark times.
-- Chris
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
