On 25-Mar-06, at 10:27 AM, Lo Saeteurn wrote:
Hmmm...I guess I'll change my animation classes to use RB
trimeshes. Wouldn't this mean that your Quesa Wrappers are
inherently slower because they are RB code?
Yes, if you hack at individual members of the "array" classes they
will be slower than using memoryblocks directly. These classes were
meant as a convenience, and are well suited for getting/submitting an
entire array but not so much for per-vertex manipulation/animation. I
purposely exposed the Trimesh Lock/Unlock (and Get/SetData) methods
so you can maintain your own data when more speed or control is
required, but here again you might need a plugin to speed up the
memoryblock manipulation.
I didn't realize memoryblocks were that much slower compared to C++
arrays.
If you're running through byte-by-byte (or vector by vector) the
function call overhead really adds up, and until RB can manipulate
memory directly it has no chance of beating or even approaching the
speed of C code in this area.
I've actually been working on a plugin with memoryblock extensions
but right now it only has a few methods to copy and count bytes -
arbitrarily sized vector math is still on the todo list.
Frank.
–––––––––––––––––––––––––––––––––
Open Source RB Goodies and Shareware
<http://developer.chaoticbox.com/>
<http://www.chaoticbox.com/>
–––––––––––––––––––––––––––––––––
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>
Search the archives of this list here:
<http://support.realsoftware.com/listarchives/lists.html>
|