www.digitalmars.com         C & C++   DMDScript  

D - glExcess into D

reply yaneurao sun-inet.or.jp writes:
glExcess is a famous visual demo used openGL.

glExcess:
http://www.glexcess.com/

I've ported glExcess into D.
http://yaneurao.zive.net/yaneSDK4D/glexcess002.zip

The latest version of 'glExcess into D' can download here:
http://www.sun-inet.or.jp/~yaneurao/dlang/english.html
(including source , opengl.d , glu.d , glut.d etc..)

yaneurao.
Jan 09 2004
next sibling parent "Walter" <walter digitalmars.com> writes:
Cool!
Jan 10 2004
prev sibling parent reply Cloud9 Virtual <cloud9virtual yahoo.com> writes:
Interesting work, I just ran both the original and the D version and 
watched the CPU load and the page file size.
The original is fast and mostly CPU bound (frequently 100%); page file 
was around 260MB

The D version is rarely CPU bound, most times is in the range of 10% 
utilisation, however it is quite slow; the page file is at around 300MB.

Since I expect the D version to be a straightforward adaptation of the 
(original C?) program, there must be a bottleneck in either your support 
library or possibly in the D compiler.
It would be interesting to know where the problem is, and what causes 
the extra 40MB of memory utilisation.


yaneurao sun-inet.or.jp wrote:
 glExcess is a famous visual demo used openGL.
 
 glExcess:
 http://www.glexcess.com/
 
 I've ported glExcess into D.
 http://yaneurao.zive.net/yaneSDK4D/glexcess002.zip
 
 The latest version of 'glExcess into D' can download here:
 http://www.sun-inet.or.jp/~yaneurao/dlang/english.html
 (including source , opengl.d , glu.d , glut.d etc..)
 
 yaneurao.
 
 

Jan 11 2004
next sibling parent reply yaneurao sun-inet.or.jp writes:
The D version is rarely CPU bound, most times is in the range of 10% 
utilisation, however it is quite slow; the page file is at around 300MB.

glexcess002.zip is a debug build because of D compiler's bug. so I change the code to avoid it. please try the latest version and give me a report again. http://yaneurao.zive.net/yaneSDK4D/glexcess003.zip yaneurao.
Jan 11 2004
parent Cloud9 Virtual <cloud9virtual yahoo.com> writes:
same behaviour. You can run the test yourself; I changed the 
fullscreen.bat parameters to match the native resolution I used with the 
native GL Excess. You may want to do trace profiling.

yaneurao sun-inet.or.jp wrote:
The D version is rarely CPU bound, most times is in the range of 10% 
utilisation, however it is quite slow; the page file is at around 300MB.

glexcess002.zip is a debug build because of D compiler's bug. so I change the code to avoid it. please try the latest version and give me a report again. http://yaneurao.zive.net/yaneSDK4D/glexcess003.zip yaneurao.

Jan 11 2004
prev sibling parent reply Ilya Minkov <minkov cs.tum.edu> writes:
I think the difference to original version is the lack of 
timer/frameskipper/soundsystem. There FMOD had been the timer.

-eye
Jan 11 2004
parent reply yaneurao sun-inet.or.jp writes:
In article <btsmv8$1mf2$1 digitaldaemon.com>, Ilya Minkov says...
I think the difference to original version is the lack of 
timer/frameskipper/soundsystem. There FMOD had been the timer.

quite true. I wasn't aware of it. I just fix it : http://yaneurao.zive.net/yaneSDK4D/glexcess004.zip the latest version can always download here: http://www.sun-inet.or.jp/~yaneurao/dlang/english.html
Jan 11 2004
parent reply Cloud9 Virtual <cloud9virtual yahoo.com> writes:
this is awesome. This version performs at the same speed and cpu usage 
as the original. On closer inspection the page file usage is very similar.
Man this is powerful stuff, as this is quite a hefty and resource hungry 
demo. It shows that even in its current alpha state the D compiler plays 
in the same league with production C/C++ compilers. And the binary is 
smaller, 65% of the original.

yaneurao sun-inet.or.jp wrote:

 In article <btsmv8$1mf2$1 digitaldaemon.com>, Ilya Minkov says...
 
I think the difference to original version is the lack of 
timer/frameskipper/soundsystem. There FMOD had been the timer.

quite true. I wasn't aware of it. I just fix it : http://yaneurao.zive.net/yaneSDK4D/glexcess004.zip the latest version can always download here: http://www.sun-inet.or.jp/~yaneurao/dlang/english.html

Jan 12 2004
next sibling parent "C" <dont respond.com> writes:
Wow thats great news!  Very exciting!  Thats a beautiful demo btw.

What sort of benchmarks exist for D , that I could use in C++ also ?

C
"Cloud9 Virtual" <cloud9virtual yahoo.com> wrote in message
news:bttmo9$blh$1 digitaldaemon.com...
 this is awesome. This version performs at the same speed and cpu usage
 as the original. On closer inspection the page file usage is very similar.
 Man this is powerful stuff, as this is quite a hefty and resource hungry
 demo. It shows that even in its current alpha state the D compiler plays
 in the same league with production C/C++ compilers. And the binary is
 smaller, 65% of the original.

 yaneurao sun-inet.or.jp wrote:

 In article <btsmv8$1mf2$1 digitaldaemon.com>, Ilya Minkov says...

I think the difference to original version is the lack of
timer/frameskipper/soundsystem. There FMOD had been the timer.

quite true. I wasn't aware of it. I just fix it : http://yaneurao.zive.net/yaneSDK4D/glexcess004.zip the latest version can always download here: http://www.sun-inet.or.jp/~yaneurao/dlang/english.html


Jan 12 2004
prev sibling parent yaneurao sun-inet.or.jp writes:
In article <bttmo9$blh$1 digitaldaemon.com>, Cloud9 Virtual says...
 It shows that even in its current alpha state the D compiler plays 
 in the same league with production C/C++ compilers.

sure. D compiler is pretty good.
 And the binary is smaller, 65% of the original.

since my 'glExcess into D' doesn't support sound and FPS(frames per second) display ,so it doesn't mean D compiler generate smaller binary than C/C++ compiler. by fairly judging , it is almost even. yaneurao.
Jan 12 2004