10 slices is nothing compare to the fixed point multiplier in an asic design. Maybe this could a design optimisation, the day you need to find some free slice, to remove this stage for few unit.
2008/2/5, André Pouliot <[EMAIL PROTECTED]>: > Kenneth Ostby wrote: > > Just my $.02 > > > > Timothy Normand Miller: > > > >> Yeah, those are hard to see. I'm guessing that the major impact is > >> coordinates. I suggest that we have two multipliers. One with > >> rounding for those cases where we need it for coordinates. And one > >> without rounding, mostly for colors. > >> > > > > Having two different multipliers for colors and coordinates seems like a > > solution that could bite us in the behind if we have to update them > > later. > > > > An interesting point would be how much we're actually gaining/loosing > > from adding a 4th step to the multiplyer? Seing how it should be > > possible to pipeline it, we shouldn't have to sacrifice a lot of > > performance to gain accuracy? > > > > > > > Well it would add a fifth stage to the multiplier. That added stage > would take something like 10 more slice in the design. Not that much but > we are already taking 35 slice and 1 multiplier. It's rather small if > you consider that we are having like 28K slice and 96 multiplier in the > spartan3-4000. > > The question is still how much space do we need for the full design? If > we have the space it's not that long to add that stage later in the > design, it would take something like 2 hour with basic simulation of the > multiplier for design correctness. > > _______________________________________________ > Open-graphics mailing list > [email protected] > http://lists.duskglow.com/mailman/listinfo/open-graphics > List service provided by Duskglow Consulting, LLC (www.duskglow.com) > _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
