Hybrid low power homogenous grapics processing units

Abstract

In an example, an apparatus comprises a plurality of execution units comprising at least a first type of execution unit and a second type of execution unit and logic, at least partially including hardware logic, to analyze a workload and assign the workload to one of the first type of execution unit or the second type of execution unit. Other embodiments are also disclosed and claimed.

1 .- 40 . (canceled) 
     
     
         41 . An apparatus comprising:
 processing resources comprising a first processing resource and a second processing resource; and   scheduling circuitry coupled to the processing resources, the scheduling circuitry to:
 analyze a workload; and 
 assign the workload to one of the first processing resource or the second processing resource. 
   
     
     
         42 . The apparatus of  claim 41 , wherein the scheduling circuitry is further to:
 receive workload resource data for the workload to be executed;   parse the workload resource data;   define a set of work group boundaries within the workload;   initiate executing of the workload at one of the first processing resource or the second processing resource, wherein the workload is assigned at least in part based on a physical proximity of the workload to the first and second processing resources or one or more characteristics of the workload; and   save a workload context when execution of the workload reaches a work group boundary in the set of group boundaries, wherein:
 the first processing resource comprises a low-performance processing resource; and 
 the second processing resource comprises a high-performance processing resource. 
   
     
     
         43 . The apparatus of  claim 42 , wherein:
 the first processing resource and the second processing resource reside on a common processor coupled to a common memory structure.   
     
     
         44 . The apparatus of  claim 42 , wherein the scheduling circuitry is further to receive hints about the one or more characteristics of the workload, wherein the hints are passed via an interface. 
     
     
         45 . The apparatus of  claim 41 , wherein the processing resources are located on a single integrated circuit. 
     
     
         46 . A method comprising:
 analyzing, by scheduling circuitry coupled to processing resources of a computing device, a workload; and   assigning, by the scheduling circuitry, the workload to one of the first processing resource or the second processing resource.   
     
     
         47 . The method of  claim 46 , further comprising:
 receiving workload resource data for the workload to be executed;   parsing the workload resource data;   defining a set of work group boundaries within the workload;   initiating executing of the workload at one of the first processing resource or the second processing resource, wherein the workload is assigned at least in part based on a physical proximity of the workload to the first and second processing resources or one or more characteristics of the workload; and   saving a workload context when execution of the workload reaches a work group boundary in the set of group boundaries, wherein:
 the first processing resource comprises a low-performance processing resource; and 
 the second processing resource comprises a high-performance processing resource. 
   
     
     
         48 . The method of  claim 47 , wherein:
 the first processing resource and the second processing resource reside on a common processor coupled to a common memory structure.   
     
     
         49 . The method of  claim 47 , further comprising receiving hints about the one or more characteristics of the workload, wherein the hints are passed via an interface. 
     
     
         50 . The method of  claim 46 , wherein the processing resources are located on a single integrated circuit. 
     
     
         51 . At least one computer-readable medium having stored thereon instructions which, when executed, cause a computing device to perform operations comprising:
 analyzing, via scheduling circuitry coupled to processing resources of the computing device, a workload; and   assigning, via the scheduling circuitry, the workload to one of the first processing resource or the second processing resource.   
     
     
         52 . The computer-readable medium of  claim 51 , wherein the operations further comprise:
 receiving workload resource data for the workload to be executed;   parsing the workload resource data;   defining a set of work group boundaries within the workload;   initiating executing of the workload at one of the first processing resource or the second processing resource, wherein the workload is assigned at least in part based on a physical proximity of the workload to the first and second processing resources or one or more characteristics of the workload; and   saving a workload context when execution of the workload reaches a work group boundary in the set of group boundaries, wherein:
 the first processing resource comprises a low-performance processing resource; and 
 the second processing resource comprises a high-performance processing resource. 
   
     
     
         53 . The computer-readable medium of  claim 52 , wherein:
 the first processing resource and the second processing resource reside on a common processor coupled to a common memory structure.   
     
     
         54 . The computer-readable medium of  claim 52 , wherein the operations further comprise receiving hints about the one or more characteristics of the workload, wherein the hints are passed via an interface. 
     
     
         55 . The computer-readable medium of  claim 51 , wherein the processing resources are located on a single integrated circuit.