Real time scheduling using expected application resource usage

A computing system architecture comprising:

• a processing resource pool including a plurality of servers;
• an interface for receiving an application to be executed;
• and
• a special purpose optimization coprocessor that receives information about the application and the processing resource pool and outputs an allocation scheme for allocating tasks of the application to the processing resource pool;
• wherein the application is analyzed to determine an expected run time use of the processing resource pool by the application, wherein the expected run time use of the processing resource pool is mapped to an optimization problem;
• and wherein the at least the portion of the processing resources is assigned using a dedicated optimization coprocessor that receives the mapping of the expected run time use of the processing resource pool and outputs a solution to the optimization problem, the solution corresponding to an assignment of the at least the portion of the processing resource pool;
• wherein analyzing the processing resource pool further includes:
• determining a plurality of latencies within each of the plurality of servers and between the plurality of servers;
• determining resource contention costs for each of the plurality of servers;
• and wherein assigning the at least the portion of the processing resources is based on the plurality of latencies and the resource contention costs.

A computer program product embodied in a non-transitory computer readable medium and comprising computer instructions for:

• receiving an application;
• analyzing the application to determine an expected run time use of a processing resource pool by the application;
• providing, to a special purpose optimization coprocessor, information about the expected run time use of the processing resource pool, the special purpose optimization coprocessor outputting an allocation scheme for allocating tasks of the application to the processing resource pool based on the expected runtime use of the processing resource pool;
• and
• assigning the tasks of the application using the allocation scheme;
• wherein the processing resources include a plurality of servers and wherein the analyzing the processing resources further includes:
• determining a plurality of latencies within each of the plurality of servers and between the plurality of servers;
• determining resource contention costs for each of the plurality of servers;
• and wherein the assigning is based on the plurality of latencies and the resource contention costs
• wherein the analyzing further includes mapping the expected run time use of the processing resources pool to an optimization problem;
• and wherein the assigning the at least the portion of the processing resources uses a dedicated optimization coprocessor that receives the mapping of the expected run time use of the processing resources and outputs a solution to the optimization problem, the solution corresponding to an assignment of the at least the portion of the processing resources.