- Find the best possible asymptotic formula for the parallel time T(n,p).
- Using this formula, calculate the cost C(n,p) and efficiency E(n,p).
- Given 0<E_{0}<1, find the asymptotically maximal function f_{1} such that
for all p_{n}=O(f_{1}(n)): E(n,p_{n})>= E_{0}.
- Given 0<E_{0}<1, find the asymptotically minimal function f_{2} such that
for all n_{p}=\Omega(f_{2}(p)): E(n_{p},p)>= E_{0}.
- Find the asymptotically minimal function f_{3} such that
for all p=\Omega(f_{3}(n)): T(n,p)=T_{opt}(n,p)
- What is the scalability of this algorithm?