On Thu, 12 Dec 2002 01:29:58 +1300, Warp wrote:

>   I successfully installed PVM for myself and compiled pvmpovray35.
> 
>   Using 12 sparcs benchmark.pov (with the suggested options) rendered in
>   34
> minutes. Not bad.

Nice to hear that it actually works for someone else :-)
It would be nice to see more parallel benchmark results on
povbench/tabsnet, so I encourage all to post their results.

Matthew.

Post a reply to this message

On Wed, 11 Dec 2002 23:14:04 +1300, Warp wrote:

>   By the way, I wouldn't like to start a sermon about programming
>   guidelines,
... 
>   For some reason many programmers have the odd idea that a header file
>   is
> a file where you declare all types and constants, regardless of where
> they are used. This is certainly not the purpose of a header file.
> 
>   Sorry, couldn't help myself... :)

I didn't write the original, I just did enough to get the thing to work
with 3.5. Quite frankly, POV-Ray and the PVM patch have enough holes to
drive a fairly large truck through.
However, I do agree that in theory, only 'public' stuff should go in the
header files, but sometimes it is not always that simple...

Matthew.

Post a reply to this message

In article <3df7e833@news.povray.org> , "Apache" 
<apa### [at] yahoocom> wrote:

> Peopla at university taught me this very thing: put types and constants in
> header file.

Which is completely wrong:  Only put types and constants used in more than
one source file into a header file.  Otherwise you just clutter the
namespace, regardless of the language being used.

    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

Agreed! I had some trouble with those people more than once about this kind
of things.

"Thorsten Froehlich" <tho### [at] trfde> wrote in message
news:3df8585a$1@news.povray.org...
> In article <3df7e833@news.povray.org> , "Apache"
> <apa### [at] yahoocom> wrote:
>
> > Peopla at university taught me this very thing: put types and constants
in
> > header file.
>
> Which is completely wrong:  Only put types and constants used in more than
> one source file into a header file.  Otherwise you just clutter the
> namespace, regardless of the language being used.
>
>     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

Apache <apa### [at] yahoocom> wrote:
> Peopla at university taught me this very thing: put types and constants in
> header file.

  Then change university. :)

-- 
#macro N(D)#if(D>99)cylinder{M()#local D=div(D,104);M().5,2pigment{rgb M()}}
N(D)#end#end#macro M()<mod(D,13)-6mod(div(D,13)8)-3,10>#end blob{
N(11117333955)N(4254934330)N(3900569407)N(7382340)N(3358)N(970)}//  - Warp -

Post a reply to this message

Warp <war### [at] tagpovrayorg> wrote:
>   Using 12 sparcs benchmark.pov (with the suggested options) rendered in 34
> minutes. Not bad.

  With 29 sparcs benchmark.pov rendered in 21 minutes.

  One problem is that it takes almost 6 minutes to parse benchmark.pov
regardless of the amount of computers, so the time will never go under
that no matter how many sparcs I use (except if I use faster sparcs,
of course).

-- 
#macro N(D)#if(D>99)cylinder{M()#local D=div(D,104);M().5,2pigment{rgb M()}}
N(D)#end#end#macro M()<mod(D,13)-6mod(div(D,13)8)-3,10>#end blob{
N(11117333955)N(4254934330)N(3900569407)N(7382340)N(3358)N(970)}//  - Warp -

Post a reply to this message

Warp wrote:
> 
>   One problem is that it takes almost 6 minutes to parse benchmark.pov
> regardless of the amount of computers, so the time will never go under
> that no matter how many sparcs I use (except if I use faster sparcs,
> of course).

Distributing the photons should be quite possible theoretically, of course
it would be quite some additional work to implement.

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 07 Dec. 2002 _____./\/^>_*_<^\/\.______

Post a reply to this message

Maybe in C, in C++ your types and constants should have the scope of the
class or namespace they are defined in, thus only cluttering that namespace,
if you can even consider that clutter.

Mike

"Thorsten Froehlich" <tho### [at] trfde> wrote in message
news:3df8585a$1@news.povray.org...
> In article <3df7e833@news.povray.org> , "Apache"
> <apa### [at] yahoocom> wrote:
>
> > Peopla at university taught me this very thing: put types and constants
in
> > header file.
>
> Which is completely wrong:  Only put types and constants used in more than
> one source file into a header file.  Otherwise you just clutter the
> namespace, regardless of the language being used.
>
>     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

Michael Goldshteyn <mgo### [at] n-o-s-p-a-m-earthlinknet> wrote:
> Maybe in C, in C++ your types and constants should have the scope of the
> class or namespace they are defined in, thus only cluttering that namespace,
> if you can even consider that clutter.

  Naturally, but even in C++ the same principle holds: Types and constants
used only inside the module should be defined in the .cpp file, not the
header file.

  Of course the exception to this in C++ are the definitions in the private
part of the (public) class, which must be declared in the same place as the
class itself (ie. in the header). Even though this is cumbersome, the reason
for this is purely technical: The compiler must know the size of the class
everywhere it's used, and member variables inside the class need to be known
in order to know this size (regardless of where they are inside the class).

  One thing which I don't like and don't agree is that in C++ you also must
declare all private methods and types in the same class definition as well.
There's no technical nor syntactical need for this. It would be perfectly
possible to declare private methods and types separately from the class
definition itself.
  For example, I don't see any problem in supporting something like this:

(In file.hh)

class Whatever
{
 public:
    void f();

 private:
    int a;
};

(In file.cc)

void Whatever::f() // Implementation of the public method f()
{
    ...
}

typedef double Whatever::D; // A private type declaration

void Whatever::g() // Declaration and implementation of a private method g()
{
    ...
}

-- 
#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

I changed classes :-)

Post a reply to this message