ravindra kalla
Junior Member level 2
hi,
i am making a architecture in which i am using processing element which used to calculate sum of absolute difference (SAD)of two matrics of size 4*4.In one case i have to calculate SAD of 18 matrics and in other case i have to calculate SAD of 10 matrics.So if i am using 9 processing element(each processing element is calculatind SAD of two matrics) in parallel for one case then in second case when i have to calculate SAD of 10 matrics then 4 processing element is remaining idle .IS thier any solution so that i can utilise hardware 100 percent for this process.
please give your valuable suggetion.
thanx
i am making a architecture in which i am using processing element which used to calculate sum of absolute difference (SAD)of two matrics of size 4*4.In one case i have to calculate SAD of 18 matrics and in other case i have to calculate SAD of 10 matrics.So if i am using 9 processing element(each processing element is calculatind SAD of two matrics) in parallel for one case then in second case when i have to calculate SAD of 10 matrics then 4 processing element is remaining idle .IS thier any solution so that i can utilise hardware 100 percent for this process.
please give your valuable suggetion.
thanx