|
|||||||
| Forum Home | Register | Members List | Mark Forums Read |
![]() |
|
|
Thread Tools | Display Modes |
|
|
#11 | ||||||||
|
Member
Location: Turin, Italy |
I personally like Haskell because of its elegance and purity. But I usually use C++ or C# to make the majority of my programs.
|
||||||||
|
|
|
|
|
#12 | ||||||||
|
Senior Member
|
How long have you guys been using C++? And lua?
|
||||||||
|
|
|
|
|
#13 | |||||||||
|
Administrator
Location: London, UK |
Quote:
__________________
Michael 'Adrir' Scott :: Games, Virtual Worlds, Education Networking | Research | Teaching |
|||||||||
|
|
|
|
|
#14 | ||||||||
|
Moderator
Location: Philadelphia, PA |
A little more than 3 years for C++. I have only looked at Lua a little.
__________________
Grant Shonkwiler() "I would love to fix the world if someone would just give me the source code" Website Industry blog LinkedIn |
||||||||
|
|
|
|
|
#15 | |||||||||
|
Junior Member
|
Quote:
![]()
__________________
E McNeill |
|||||||||
|
|
|
|
|
#16 | ||||||||
|
Member
Location: Helsinki, Finland |
Umm, actually I do, but thats beside the point :P
Small costs here and there form a huge river. Array index check here. Garbage collection there. Just for the record, I recently wrote a basic raycaster in java just for the fun of it. Nothing fancy. The everyday symmetrical grid based thingy. Decided to use Java2D trusting that it might just be effective enough... Lo and behold. Profiling reveals that 95% of the execution calls consists of raster array manipulations, scaleImages and drawImages. (Managed to eliminate scaleImage calls alltogether by somesort of caching class, but that did not do much). The thing works nicely in resolutions under 640 x 480. However, squeezing any more than 15 fps from resolutions above that mark proved difficult. The walls render well enough, but the darn ceiling and floor casting takes insane amounts of raster.setElem(indx, color); Not only is it a method call, but most certainly there is also an array indexing happening inside the thing. So here I am paying for: method call spiced with passing of 2 params. + possible unwanted array index checkup Doubt the memory cache likes this type of foolery very much. At least the fps stays ridiculously low with hundreds of thousands of individual bloated operations. go Java! ![]() Last edited by DTR : 03-01-2009 at 03:30 AM. |
||||||||
|
|
|
|
|
#17 | ||||||||
|
Administrator
Location: UK |
That's a problem with the library, not the language.
__________________
Steven Yau [Alix Games Blog] [Portfolio] [How I broke into the Games Industry] [Why I left my Games Job] [How to be a Games Tester] [Getting back into the Game] |
||||||||
|
|
|
|
|
#18 | ||||||||
|
Member
Location: Helsinki, Finland |
That may be so.
Im currently making the thing make use of OpenGL with JOGL instead. Its not like the array indexing or method calling is going to be significantly less with it. The fact being that huge majority of time went on floor and ceiling rendering. (drawing the horizontal resolution amount of differently scaled slices was not particulary heavy task) I wrote the pixels directly into the raster, which I drew at the end of the rendering on screen. The actual raycasting operation took less than 7 % of the total execution time and I was happy with that. That one blitting at the end of the rendering just cant have been that consuming, unlike the constant array operations assigning pixel values. I just have a feeling that the performance is not going to increase too much with OpenGL.... But should it do so, well. I would ge glad. |
||||||||
|
|
|
|
|
#19 | ||||||||
|
Administrator
Location: UK |
Can't you turn off the array bounds check when you compile it?
__________________
Steven Yau [Alix Games Blog] [Portfolio] [How I broke into the Games Industry] [Why I left my Games Job] [How to be a Games Tester] [Getting back into the Game] |
||||||||
|
|
|
|
|
#20 | ||||||||
|
Member
Location: Helsinki, Finland |
I have read a few sources on the internet about the thingy, and it definately does not seem practical.
http://c2.com/cgi/wiki?CatchDontCheck Suggestions range from abusing try{}catch() to using array.length as the loops cap value have been suggested. For me it seems as if the thing works much like the garbage collector. If works if the compiler so pleases without asking many questions. Only reason I chose java for the task anyway was the crossplatfom compatibility and the apparently faster prodution speed. |
||||||||
|
|
|
![]() |
«
Previous Thread
|
Next Thread
»
| Thread Tools | |
| Display Modes | |
|
|
Powered by vBulletin® Version 3.6.9
Copyright ©2000 - 2013, Jelsoft Enterprises Ltd.
Copyright ©2000 - 2013, Jelsoft Enterprises Ltd.
All times are GMT -8. The time now is 11:22 AM.































Linear Mode

