Has anyone tried to compile 3.5 with 64bit optimizations?  I have a new PC 
with an AMD64 3200+ cpu (running Windows XP64) and I'm dying to see what 
this thing can do in native 64bit mode.

Rich

Post a reply to this message

Rich wrote:
> Has anyone tried to compile 3.5 with 64bit optimizations?  I have a new PC 
> with an AMD64 3200+ cpu (running Windows XP64) and I'm dying to see what 
> this thing can do in native 64bit mode.

Correct me if i am wrong but AFAIK there isn't any 64 bit Windows 
version available at the moment (so what you have probably either is a 
beta version or a marketing gag).

In general POV-Ray compiles on 64bit systems but it won't be faster (in 
most cases even slower).  The main advantage would be to be able to 
address more than 4GB of memory.

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 21 Mar. 2004 _____./\/^>_*_<^\/\.______

Post a reply to this message

Rich <SrP### [at] ricoswebcom> wrote:
> I'm dying to see what this thing can do in native 64bit mode.

  What do you expect it to do that it can't do right now? Or what do you
expect it to do better than now?

  Unless you have more than 4 gigabytes of memory (which I doubt), you
will not probably see any difference (and even if you have more memory
than that, you will see the difference only if you make a scene which
consumes that much memory). If anything, the 64-bit version will be
some percents slower (except in the AMD 64-bit processor, where they
for some odd reason did not optimize 32-bit execution as much as the
64-bit execution like every other 64-bit processor has done).

  If you expect POV-Ray to do something better when compiled to 64-bit
architecture, could you explain what and why?

-- 
#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

Warp <war### [at] tagpovrayorg> wrote:
> some percents slower (except in the AMD 64-bit processor, where they
> for some odd reason did not optimize 32-bit execution as much as the
> 64-bit execution like every other 64-bit processor has done).

Well, he did say that he has an AMD64, so he probably only cares about 
that, not about 64bit in general.

And AFAIK the AMD64 has more registers in 64bit mode than in 32bit,
mode, wich isn't odd, since they won't be of any use in 32bit mode anyways
(both the OS and software need to "know" about them to take advantage).

Every time someone here asks about an AMD64 version they get replies
that 64bit versions in general won't be of much use, but it seems so far
nobody has ACTUALLY compiled it for AMD64 and tried it... *sigh*

Lutz-Peter

Post a reply to this message

Lutz-Peter Hooge wrote:
> [...]
> 
> Every time someone here asks about an AMD64 version they get replies
> that 64bit versions in general won't be of much use, but it seems so far
> nobody has ACTUALLY compiled it for AMD64 and tried it... *sigh*

If you give me an AMD 64bit system i will try it right away...

But your "Every time someone here asks" is completely exaggerated - i 
think no one here has seriously offered to try running POV on such a system.

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 21 Mar. 2004 _____./\/^>_*_<^\/\.______

Post a reply to this message

Lutz-Peter Hooge <lpv### [at] gmxde> wrote:
> Well, he did say that he has an AMD64, so he probably only cares about 
> that, not about 64bit in general.

  If he has an AMD64 then he will probably benefit from a 64-bit version
of POV-Ray for the reason I explained before: AFAIK for some reason 32-bit
code does not run as fast as 64-bit code in this processor. (But this has
nothing to do with 64-bit code per se being generally faster than 32-bit
code, but because of the optimization strategies of the processor.)

  I find it odd that AMD made this decision. Other examples of 32/64 bits
processors are the Sun UltraSparc and the G5 processor of Macintoshes.
In both of them the processor has been fully optimized for both 32-bit and
64-bit code execution and there isn't any drastic difference between the
two (even though 64-bit code is generally a few percents slower because
the code takes more space and fills caches a bit faster).

  Of course what AMD did is much much better than what Intel did...

> Every time someone here asks about an AMD64 version they get replies
> that 64bit versions in general won't be of much use, but it seems so far
> nobody has ACTUALLY compiled it for AMD64 and tried it... *sigh*

  What I choke at is when people talk as if 64-bit code would have something
magical which would make POV-Ray drastically faster. I just asked exactly
what these people are expecting. Why would it be faster?

-- 
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

Warp <war### [at] tagpovrayorg> wrote:

> I find it odd that AMD made this decision. 

I guess because they want people to actually buy their processors, so
they shouldn't get too expensive?

> I just asked exactly what these people are expecting. Why would it be faster?

Of course he expects it to be faster. Why? Because he has not an "generic" 
64bit processor but a special implementation wich can very well be faster in
64bit mode.

Lutz-Peter

Post a reply to this message

In article <4082906d@news.povray.org> , Warp <war### [at] tagpovrayorg>  wrote:

>  AFAIK for some reason 32-bit
> code does not run as fast as 64-bit code in this processor.

Well, the 64-bit extension is essentially a RISC design except for the
instruction encoding.  And you get twice as many registers as in 32-bit
mode.  Not really difficult to optimise for such a design, nor hard to
implement an efficient core based on a clean RISC archicture ... in
particular with just 16 registers.

    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

Yes, I am using the beta version of Windows XP 2003 64bit, available for 
free from Microsoft to anyone willing to download it (or pay a small fee 
to recieve it on CD).

From what I've heard on the WindowsXP64 board, compiling with the 
optimizations intended for 64bit, or possibly for the AMD64 particularly, 
will increase program execution speed dramatically.  That's what I 
expect, along with a better/faster/more efficient numerical precision 
when using really small, or really large, parameters in object size, 
placement, etc.

I am a web developer using c# and asp...  I don't know much at all about 
compiler optimizations other than what I've read.  What's being written 
is that a program compiled to 32bit will run at simliar speeds on an AMD 
XP and an AMD64 rated similarly.  When that same program is compiled to 
64bit, using the optimizations intended for 64bit, the program will run 
significantly faster on the 64bit processor (and not at all on the 32bit 
processor).

If I'm wrong, then so are a lot of developers posting on the WindowsXP64 
board.  <shrug>  I'm already getting a big boost over the Athlon XP 1800+ 
I was running POV on before, so I'm very satisfied with my upgrade.  If 
there's even a slight possiblity that compiling with 64bit optimizations 
will result in faster code execution, why wouldn't I want to try it?

As soon as .Net is able to be run on Windows XP 64bit, I'd be willing to 
attempt the compile myself (hopefully with help from this group! :).  If 
there's a freeware compiler that can do 64bit executables on Windows XP 
64bit, I'd be willing to try that route also.

Rich

Post a reply to this message

Rich wrote:
> Yes, I am using the beta version of Windows XP 2003 64bit, available for 
> free from Microsoft to anyone willing to download it (or pay a small fee 
> to recieve it on CD).
> 
> [...]
> 
> As soon as .Net is able to be run on Windows XP 64bit, I'd be willing to 
> attempt the compile myself (hopefully with help from this group! :).  If 
> there's a freeware compiler that can do 64bit executables on Windows XP 
> 64bit, I'd be willing to try that route also.

Do i understand this right, MS offers a free beta version but not a 
compiler suited for generating the appropriate programs (neither free 
nor for $$$)?

You can also have it easier, AFAIK both Linux and FreeBSD are available 
for 64bit AMD.

> From what I've heard on the WindowsXP64 board, compiling with the 
> optimizations intended for 64bit, or possibly for the AMD64 particularly, 
> will increase program execution speed dramatically.  That's what I 
> expect, along with a better/faster/more efficient numerical precision 
> when using really small, or really large, parameters in object size, 
> placement, etc.

Yes, and it will also increase you car's speed, make cows give more milk 
and leads to more sunny days...

SCNR.

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 21 Mar. 2004 _____./\/^>_*_<^\/\.______

Post a reply to this message