Asymmetric heterogeneous multi-threaded operating system

• US 7,516,456 B2
• Filed: 09/25/2003
• Issued: 04/07/2009
• Est. Priority Date: 09/25/2003
• Status: Expired due to Fees

• Current Assignee: GlobalFoundries, Inc.
• Original Assignee: International Business Machines Corporation
• Inventors: Nutter, Mark Richard, Day, Michael Norman, Stafford, James Michael, Aguilar, Maximino Jr.
• Primary Examiner(s): Lee; Thomas
• Assistant Examiner(s): WAI, ERIC CHARLES
• Application Number:
  US10/670,841
• Publication Number:
  US 20050081203A1
• Time in Patent Office: 2,021 Days
• Field of Search: 718102-103, 711/120
• US Class Current: 718/102
• CPC Class Codes: G06F 9/4881 Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues

Abstract

A method for an asymmetric heterogeneous multi-threaded operating system is presented. A processing unit (PU) provides a trusted mode environment in which an operating system executes. A heterogeneous processor environment includes a synergistic processing unit (SPU) that does not provide trusted mode capabilities. The PU operating system uses two separate and distinct schedulers which are a PU scheduler and an SPU scheduler to schedule tasks on a PU and an SPU, respectively. In one embodiment, the heterogeneous processor environment includes a plurality of SPUs. In this embodiment, the SPU scheduler may use a single SPU run queue to schedule tasks for the plurality of SPUs or, the SPU scheduler may use a plurality of run queues to schedule SPU tasks whereby each of the run queues correspond to a particular SPU.

Claims

1. A computer implemented method for asymmetrically multithreading tasks in a computer environment that includes a plurality of dissimilar processors, said method comprising:

retrieving an application;

analyzing a first task type and a second task type included in the application;

in response to analyzing the first task type and the second task type, determining that the application requires both a first scheduler and a second scheduler,

wherein the first scheduler is different than the second scheduler;

scheduling, during execution of the application, a plurality of first tasks adapted to be executed by a first type of processor from the plurality of dissimilar processors, the scheduling performed by the first scheduler that maintains a first run queue that includes data corresponding to the first tasks, the plurality of first tasks corresponding to the first task type;

in response to the scheduling of the plurality of first tasks by the first scheduler,

inserting the plurality of first tasks in the first run queue;

scheduling, during the execution of the application, a plurality of second tasks adapted to be executed by a second type of processor from the plurality of dissimilar processors, the scheduling performed by the second scheduler that maintains a second run queue that includes data corresponding to the second tasks, wherein the first scheduler and the second scheduler are both utilized by the application during the application execution, and wherein the first task and the second task execute concurrently, the plurality of second tasks corresponding to the second task type;

in response to the scheduling of the plurality of second tasks by the second scheduler, inserting the plurality of second tasks in the second run queue;

informing the second type of processor to load one of the plurality of second tasks in response to the scheduling of the plurality of second tasks; and

wherein the first scheduler and the second scheduler are both located on a first processor that is the first type of processor.

2. The method as described in claim 1 wherein the scheduling operation of the plurality of first tasks is asymmetric to the scheduling operation of the plurality of second tasks.

3. The method as described in claim 1 further comprising:

wherein a first identifier space corresponds to the first type of processor and

wherein a second identifier space corresponds to the second type of processor; and

wherein a first task list corresponds to the first type of processor and wherein a second task list corresponds to the second type of processor.

4. The method as described in claim 1 wherein the scheduling operation of the plurality second tasks further comprises:

receiving a new task from the plurality of second tasks;

identifying new task attributes corresponding to the new task;

comparing the new task attributes with one or more scheduled task attributes, the scheduled task attributes corresponding to one or more scheduled tasks that are included in the second run queue; and

performing the scheduling of the new task based upon the comparing.

5. The method as described in claim 4 wherein at least one of the new task attributes are selected from the group consisting of a policy and a priority.

6. The method as described in claim 1 wherein the computer environment includes a plurality of second type of processors, and wherein the second scheduler maintains a plurality of second run queues, each of the plurality of second run queues corresponding to each of the plurality of second type of processors.

7. The method as described in claim 6 wherein the scheduling operation of the plurality of second tasks further comprises:

receiving a new task from the plurality of second tasks analyzing a plurality of workloads that correspond to the plurality of second type of processors;

identifying an available second type of processor from the plurality of second type of processors in response to the analyzing; and

including the new task in the second run queue from the plurality of second run queues that corresponds to the available second type of processor.

8. The method as described in claim 1 wherein the first type of processor is a processing unit that operates in a trusted mode environment and wherein the second type of processor is a synergistic processing unit that operates in a non-trusted mode environment.