Locality-Aware Task Scheduling and Data Distribution for OpenMP Programs on NUMA Systems and Manycore Processors

Algorithm 3

Work-dealing algorithm for TILEPro64.

(1)  Procedure  deal-work(task T, queues , current home cache n,

  current data distribution policy p, access-intensive dependence index a)

(2)  if    then

(3)     if  exists a  then

(4)       find  containing T.depend_list[];

(5)       enqueue(, T);

(6)     else

(7)       Populate with bytes in T.depend_list;

(8)       if    then

(9)       find with least home cache latency cost to ;

(10)        enqueue(, T);

(11)         else

(12)         enqueue(, T);

(13)         end

(14)     end

(15)    else

(16)     enqueue(, T);

(17)    end

(18) end