On Wed, 2023-03-29 at 10:30 -0700, Nick Desaulniers wrote: > > > > > Nick pointed out that the result of this division is being > > > stored to a 32-bit type anyways, so truncate gp2_10ns first then do the > > > division, which elides the need for libcalls. > > > > That loses ~7 top bits though, no? I'd be more worried about that, than > > the time div_u64() takes. > > The result is still stored in a u32; there is a loss of precision > regardless of use of div_u64 or open coded binary operator /. > Right, obviously. > So is > the loss of precision before the division as tolerable as after the > division? For all I can tell this is meant to be 'gp2' with an additional lower bits to reach a unit/granularity of 10ns, basically in FW something like gp2_10ns = gp2 * 100 + subsampling_10ns_unit (and gp2 in FW is a 32-bit value, so it rolls over eventually). But I _think_ we want to make a proper division by 100 to obtain back the original 'gp2' value here. johannes
