D3DXMATRIXA16 what "A16" mean?

Started by
6 comments, last by katsh 17 years, 6 months ago
D3DXMATRIXA16 , -A16 is what? my guess is,, 1. 16bit Alignment 2. cause there is 16 element(_11..._44) which 1? or other meaning?
Advertisement
1.
The SDK documentation explains it rather well, and in much more detail. But yes, it's the alignment option.
thx guys.
smoke in my brain was wiped.
Aligned on 16bit memory
16 *byte* aligned.
do unto others... and then run like hell.
Quote:Original post by FReY
16 *byte* aligned.
Yup.

The reason for the alignment is that it can be faster to use the matrix if it's aligned to a certain boundary. SSE2 and 3dNow! both need the matrix aligned. If it's not aligned, I'd guess that D3DX will either copy it to an aligned version, or it'll fallback and use some other method for changing the matrix (E.g. a plain x86 ASM version of a matrix multiply rather than an SSE2 one).

Basically, it's faster to use, but usually takes more memory.
i see.thx for the information.
i realize now that why D3DX can calculate matrices so fast.

This topic is closed to new replies.

Advertisement