Oops, screwed up..
The previous frame rates were wrong...
I compared my routines (using e_mem_copy) with se_mem_copy and was told
that se_mem_copy went about 2 times faster (319 fps). This suprised me,
cause it was against all logic.
But what happend..
I copied a whole screen of 320*200 to #A0000 using se_mem_copy
And because of a bug in my code, i copied a whole screen of 320*200
to #A0000 + 1. So 1 byte would be out of the video memory... this made
the framerate drop to 180 fps.
I really don't understand this, but i can tell you that, now.. with
the bug out I get these results:
- 346 fps on my P-60 (with win95 running, which slows things down)
And se_mem_copy gave 316 fps (also on the same computer)
So my routines (using command list --> e_mem_copy ) go 8-9% faster
than se_mem_copy.
Still, my routines will give you the same flexibility of video modes
and screens and sprites, etc.
I now promise to stop posting about nextGFX, i know this looks kind
of arogant, however i told you people this, because i was curious why,
and because some of you (like me) would be shocked when they heard that
the framerate was only 180 fps. I told you about the framerates in the
first mail, because i was suprised and still am curious, what the
difference between source and dest. copying actually is.
Ralf Nieuwenhuijsen
nieuwen at xs4all.nl
|
Not Categorized, Please Help
|
|
