id	summary	reporter	owner	description	type	status	priority	milestone	component	version	resolution	keywords	cc
655	BDOF multiplication of sGxGy  and vx	bheng		"Is there a reason sGxGy is split into Hi/Lo bits in eq 8-806 and 8-807?

   sGxGym = sGxGy >> 12
   sGxGys =  sGxGy & ( ( 1 << 12 ) − 1 )

The two halves are immediately put back together after the multiplication with vx below:

   ( ( vx * sGxGym )  <<  12 + vx * sGxGys )  

And that is the only place these variables are used.

Isn't this equivalent to just (vx * sGxGy), without needing all these extra steps?

"	defect	closed	minor		spec	VVC D7 vB	fixed		ksuehring bbross XiangLi fbossen jvet@…