Hollywood forums

Sic... How we can accelerate our hollywood game?! It's not calcul, it's just display problem.

I think with 460 we can make game with multidirectionnal scrolling in 800*600 but in 640*480 32bits, my scroll engine limit is 30FPS.

So when i use tile technique, the engine have the same speed that displaying 640*480 picture. Displaybrush (with tiles of 32x32 pixel) is equivalent in speed that displaybrushpart (with picture 640x480 pixel)... Normaly, tiles technique is speeder than displaying big picture.

Other thing, doublebuffer is slower than my createbrush technique, i win 4FPS without doublebuffer on my scroll (on microAone), or doublebuffer is normally the speedest animation technique and other, if i use a lot of function in the same programm, the code is slower.

try this:

http://jeannoel.roueste.free.fr/Amiga/S ... ngine2_OS4

On my Peg2@1131 + Radeon9000pro i gives me 67/71 fps and was really nice to play

However I think that the only way we have to get better performances with Hollywood was to have a JIT player rather than a complied source...maybe easyer to do for Andreas and better for we to handle...

PS I get a littel bug randomly occured on your engine... on about 5 games I get 2 time tha terror: "Table field 60 was not initialiesed!" in line 413 in function concurrent...

keep up good work!

My IA engine is bad and the game is too slow, so, i wait new version of hollywood, and i pray for new 2d engine... at this time, it's not interessant to develop game (with scrolling) with holly for me, i just create some little program for the fun

Andreas wrote:It could also be that fetching Intuition events is slower on OS4 than on MorphOS. Hollywood fetches these events quite often to stay fully responsive. If OS4 is slower here than MorphOS this would be an explanation for the significant speed differences between both systems.

There was any way to test that? I mean... is there any specific benchmark to test what the problem can be? To isolate the problem...than I try to report it on the AmigaOS4.x support forum and see what they think of it...

Thank you!

I found this:

http://luajit.org/index.html

I think it's can help for futur dev of hollywood

ArtBlink wrote:I found this:

http://luajit.org/index.html

I think it's can help for futur dev of hollywood

Well...only Andreas can say that...but hope that!

However atm I hope to get where and why MOS was faster than OS4.x and try to solve that...than a JIT will be REALLY appreciated!

Tuxedo wrote: There was any way to test that? I mean... is there any specific benchmark to test what the problem can be?

I don't have any but it wouldn't be hard to write one.

I Can do it with Hollywood or must be wrote in C ?

No, this would have to be done in C or Assembler.