Re: [eigen] Rigid Transformations in eigen: discussion thread

[ Thread Index | Date Index | More Archives ]

2009/9/17 Rohit Garg <rpg.314@xxxxxxxxx>:
> If you are talking about AVX, then AFAIK, it won't need 32 byte
> alignment. It'll happily work with 16 byte aligned pointers.

ah OK, but perhaps we shouldn't rely on such hardware details, if we
want future-proof code!

Fortunately, this can be changed at a later date without affecting the
API and ABI.

Still, i am thinking that it is slightly more future proof to allocate
a vector8 and map as quaternions, than the inverse.


Mail converted by MHonArc 2.6.19+