Siarhei Siamashka siarhei.siamas...@gmail.com writes:
The main culprits are functions pixman_transform_point_3d() and
pixman_transform_point() here:
http://cgit.freedesktop.org/pixman/tree/pixman/pixman-matrix.c?id=pixman-0.28.0#n49
They perform multiplication of a matrix (16.16 fixed
On Thu, 06 Dec 2012 14:02:11 +0100
sandm...@cs.au.dk (Søren Sandmann) wrote:
Søren Sandmann sandm...@cs.au.dk writes:
Intermediate results from matrix and vector multiplications fit in 64
bits, and
That is, as long as the final result doesn't overflow the 31.16 type,
the intermediate