Multi-threaded processor with deferred thread output control

• US 8,869,147 B2
• Filed: 05/31/2006
• Issued: 10/21/2014
• Est. Priority Date: 05/31/2006
• Status: Active Grant

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: Du, Yun, Jiao, Guofang, Yu, Chun
• Primary Examiner(s): Puente, Emerson
• Assistant Examiner(s): Labud, Jonathan R
• Application Number:
  US11/445,100
• Publication Number:
  US 20070283356A1
• Time in Patent Office: 3,065 Days
• Field of Search: None
• US Class Current: 718/100
• CPC Class Codes: G06F 9/30123 according to context, e.g. thread buffers
• CPC Class Codes: G06F 9/3836 Instruction issuing, e.g. dynamic instruction scheduling or out of order instruction execution
• CPC Class Codes: G06F 9/3851 from multiple instruction streams, e.g. multistreaming
• CPC Class Codes: G06F 9/3856 Reordering of instructions, e.g. using queues or age tags
• CPC Class Codes: G06F 9/3858 Result writeback, i.e. updating the architectural state or memory
• CPC Class Codes: G06F 9/4881 Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

A multi-threaded processor is provided that internally reorders output threads thereby avoiding the need for an external output reorder buffer. The multi-threaded processor writes its thread results back to an internal memory buffer to guarantee that thread results are outputted in the same order in which the threads are received. A thread scheduler within the multi-threaded processor manages thread ordering control to avoid the need for an external reorder buffer. A compiler for the multi-threaded processor converts instructions that would normally send processed results directly to an external reorder buffer so that the processed thread results are instead sent to the internal memory buffer of the multi-threaded processor.

Claims

1. A multi-threaded processor comprising:

a thread scheduler configured to track a sequence in which a plurality of threads is received at an input interface of the multi-threaded processor;

an internal memory buffer configured to temporarily store thread results corresponding to the plurality of received threads; and

a processing unit coupled to the thread scheduler and the internal memory buffer, the processing unit configured to;

process the plurality of threads out of sequence to obtain a plurality of corresponding thread results, and

store the plurality of thread results in the internal memory buffer,

wherein the thread scheduler is configured to cause the plurality of thread results stored in the internal memory buffer to be outputted from the multi-threaded processor in the sequence in which the corresponding threads were received by the multi-threaded processor.

2. The multi-threaded processor of claim 1 wherein the plurality of threads are processed by the processing unit according to the order defined by flow control instructions associated with the plurality of threads.

3. The multi-threaded processor of claim 2 wherein the flow control instructions are configured to cause the plurality of threads to be processed in a different sequence than they were received.

4. The multi-threaded processor of claim 1 wherein the internal memory buffer further includes:

a plurality of input registers configured to store the plurality of received threads prior to processing and a plurality of output registers configured to store the plurality of thread results prior to being outputted.

5. The multi-threaded processor of claim 4 wherein the processing unit is further configured to retrieve a thread from one of the plurality of input registers in the memory buffer.

6. The multi-threaded processor of claim 1 further comprising:

the input interface, wherein the input interface is coupled to the thread scheduler and configured to receive the plurality of threads; and

an output interface coupled to the memory buffer from which the plurality of stored results are outputted, wherein the output interface comprises a demultiplexer.

7. The multi-threaded processor of claim 1 further comprising:

a load controller coupled to the thread scheduler and configured to store the plurality of threads in a plurality of input registers in the internal memory buffer under the direction of the thread scheduler.

8. The multi-threaded processor of claim 7 wherein the load controller is configured to cause the results to be output from the internal memory buffer under the direction of the thread scheduler.

9. The multi-threaded processor of claim 1 wherein the received threads include pixel data and wherein the multi-threaded processor is configured to obtain data associated with a selected thread of the plurality of threads from a texture engine.

10. A multi-threaded processor comprising:

means for tracking a sequence in which a plurality of threads is received at an input interface of the multi-threaded processor;

means for processing the plurality of threads out of sequence to obtain a plurality of corresponding results;

means for storing the plurality of results in an internal memory buffer; and

means for causing the plurality of stored results to be outputted from the internal memory buffer in the sequence in which the corresponding threads were received by the multi-threaded processor.

11. The multi-threaded processor of claim 10 wherein the plurality of threads are processed according to the order defined by flow control instructions associated with the plurality of threads.

12. The multi-threaded processor of claim 10 further comprising:

means for storing the plurality of threads in the internal memory buffer prior to processing.

13. A method for reordering the sequence of a plurality of thread results within a multi-threaded processor, comprising:

tracking a sequence in which a plurality of threads is received by the multi-threaded processor;

processing the plurality of threads out of sequence on a single core to obtain a plurality of corresponding results;

storing the plurality of results in an internal memory buffer; and

sending out the plurality of stored results stored in the internal memory buffer from the multi-threaded processor in the sequence in which the corresponding threads were received by the multi-threaded processor.

14. The method of claim 13 wherein the plurality of threads are processed according to the order defined by flow control instructions associated with the plurality of threads.

15. The method of claim 13 further comprising:

receiving a plurality of threads for a particular process at a multi-threaded processor; and

storing the plurality of threads in the memory buffer prior to processing.

16. A graphics processor comprising:

a multi-threaded processor having a core, the core configured to;

track a sequence in which a plurality of threads including pixel data is received from a first application;

store the plurality of received threads in an internal memory buffer;

process the plurality of threads in a different sequence than the sequence in which the plurality of threads were received according to an order defined by flow control instructions associated with the plurality of threads to obtain a plurality of corresponding results;

store the plurality of results in the internal memory buffer; and

output the plurality of results to the first application from the internal memory buffer in the sequence in which the corresponding threads were received from the application.

17. The graphics processor of claim 16 further comprising a texture engine in communication with the multi-threaded processor, wherein the core comprises a shader core configured to perform shading operations and obtain texture data associated with a selected thread of the plurality of threads from the texture engine.

18. A non-transitory computer-readable medium for use in reordering the sequence of a plurality thread results within a multi-threaded processor, the medium comprising instructions stored thereon to cause the multi-threaded processor to:

track a sequence in which a plurality of threads is received by the multi-threaded processor;

process the plurality of threads out of sequence on a single core to obtain a plurality of corresponding results;

store the plurality of results in an internal memory buffer; and

send out the plurality of stored results from the internal memory buffer in the sequence in which the corresponding threads were received by the multi-threaded processor.

19. The non-transitory computer-readable medium of claim 18, wherein the plurality of threads are processed according to an order defined by flow control instructions associated with the plurality of threads.

20. The non-transitory computer-readable medium of claim 18, further comprising instructions to cause a processor to:

receive a plurality of threads for a particular process at an input interface of the multi-threaded processor; and

store the plurality of threads in the internal memory buffer prior to processing.