Re: [AD] About the speed of flat polygon primitives

[ Thread Index | Date Index | More lists.liballeg.org/allegro-developers Archives ]


Hi all

I've been away for a few days and when I came back I found out latest Allegro CVS won't compile on 
BeOS... I read the latest emails and it happens you all applied some modifications that made the port 
broken. Now, I'm about to apply a patch that fixes the BeOS fullscreen driver to use the new API for the 
fetch_mode_list() function, plus I tested the hardware accelerated vs software mode thingie. Here are 
the results from doing "cmplog accel.log noaccel.log":

DRAW_MODE_SOLID results:

	putpixel()                       = 234%
	hline()                          = 368%
	vline()                          = 91%
	line()                           = 92%
	rectfill()                       = 14%
	circle()                         = 103%
	circlefill()                     = 174%
	ellipse()                        = 82%
	ellipsefill()                    = 204%
	arc()                            = 88%
	triangle()                       = 206%

Other functions:

	textout()                        = 169%
	vram->vram blit()                = 1%
	aligned vram->vram blit()        = 1%
	blit() from memory               = 104%
	aligned blit() from memory       = 106%
	vram->vram masked_blit()         = N/A
	masked_blit() from memory        = 104%
	draw_sprite()                    = 113%
	draw_rle_sprite()                = 113%
	draw_compiled_sprite()           = 113%
	draw_trans_sprite()              = 100%
	draw_trans_rle_sprite()          = 100%
	draw_lit_sprite()                = 112%
	draw_lit_rle_sprite()            = 113%

Get your conclusions here...
PS: thanks to this test I discovered the software only fullscreen driver on BeOS was broken due to a 
little mistake of mine. This has been fixed.
PPS: some of the modifications made recently broke the lib building process: now it can't find obj/beos/
asmdef.inc, and I had to create it manually using "make obj/beos/asmdef.inc"
PPPS: for some strange reason now make attempts to create the bfixicon tool (BeOS specific) object 
as obj/beos/alleg/beos/bfixicon.o: that directory obviously doesn't exist! And last time I checked it 
compiled fine by placing the .o into obj/beos/alleg...

--
Angelo Mottola
a.mottola@xxxxxxxxxx
http://www.ecplusplus.com 



Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/