Method and apparatus for dynamic allocation of multiple buffers in a processor

• US 5,778,245 A
• Filed: 03/01/1994
• Issued: 07/07/1998
• Est. Priority Date: 03/01/1994
• Status: Expired due to Term

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Gupta, Shantanu R., Griffith, James S., Glew, Andrew F., Fetterman, Michael A., Hinton, Glenn J., Colwell, Robert P., Papworth, David B.
• Primary Examiner(s): Donaghue, Larry D.
• Application Number:
  US08/204,861
• Time in Patent Office: 1,589 Days
• Field of Search: 395/800, 395/375, 395/800.23, 395/800.24, 395/388, 395/394, 395/391, 395/392, 395/393, 395/563, 395/382
• US Class Current: 712/23
• CPC Class Codes: G06F 9/3836 Instruction issuing, e.g. dynamic instruction scheduling or out of order instruction execution
• CPC Class Codes: G06F 9/384 Register renaming
• CPC Class Codes: G06F 9/3854 Instruction completion, e.g. retiring, committing or graduating
• 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 9/3885 using a plurality of independent parallel functional units
• CPC Class Codes: G06F 9/5016 the resource being the memory

Abstract

A method and apparatus for dynamically allocating entries of microprocessor resources to particular instructions in an efficient manner to efficiently utilize buffer size and resources. The pipelined and superscalar microprocessor is capable of speculatively executing instructions and also out-of-order processing. Resources within the microprocessor include a store buffer, a load buffer, a reorder buffer and a reservation station. The reorder buffer contains a larger set of physical registers and also contains information related to speculative instructions and the reservation station comprises information related to instructions pending execution. The load buffer is only allocated to load instructions and is valid for an instruction from allocation pipestage to instruction retirement. The store buffer is only allocated to store instructions and is valid for an instruction from allocation to store performance. The reservation station is allocated to most instructions and is valid for an instruction from allocation to instruction dispatch. The reorder buffer is allocated to all instructions and is valid for a given instruction from allocation to retirement. The load buffer, store buffer, and reorder buffer are sequentially allocated while the reservation station is not. Resource allocation is performed dynamically (as needed by the operation) rather than as a full set of resources attached to each operation. Using the above allocation scheme, efficient usage of the microprocessor resources is accomplished.

Claims

1. A pipelined processor comprising:

a reorder buffer containing result information generated by speculative instructions;

a reservation station coupled to said reorder buffer, said reservation station containing information associated with instructions pending execution; and

an allocator coupled to said reorder buffer and said reservation station, said allocator allocating entries of said reorder buffer and said reservation station during an allocation pipestage of said pipelined processor, said allocation pipestage following a decode pipestage of said pipelined processor and preceding a dispatch pipestage of said pipelined processor, said allocator locating entries of said reservation station for each pending instruction except for one or more register exchange instructions.

2. The processor of claim 1 further comprising:

a load buffer containing information associated with load instructions;

a store buffer containing information associated with store instructions; and

wherein said allocator allocates entries of said load buffer to said load instructions during said allocation pipestage and allocates entries of said store buffer to said store instructions during said allocation pipestage.

3. The processor of claim 2 wherein said allocator comprises:

sequential allocation means for sequentially allocating entries of said reorder buffer, said load buffer and said store buffer; and

nonsequential allocation means for nonsequentially allocating entries of said reservation station.

4. The processor of claim 3 wherein said reorder buffer, said load buffer and said store buffer each comprise a circular buffer having a tail pointer indicating a last valid entry and a head pointer indicating a first valid entry.

5. The processor of claim 2 wherein said reorder buffer includes a plurality of physical registers for containing said result information and said reorder buffer expands a register set associated with said processor.

6. The processor of claim 2 wherein said allocator comprises instruction allocation means for allocating a full set of instructions to said reorder buffer, a subset of instructions to said reservation station, load instructions to said load buffer and store instructions to said store buffer.

7. The processor of claim 2 further comprising means for stalling said processor from issuing instructions provided entries are unavailable within said reorder buffer, said reservation station, or said load buffer.

8. The processor of claim 2 further comprising a bus coupled to said processor and wherein said store instructions are executed when committed to memory over said bus.

9. A computer system comprising:

a bus for communicating information;

a memory storage unit coupled to said bus for storing information;

a first processor coupled to said bus; and

a second processor coupled to the bus, the second processor being a pipelined processor comprising;

a reorder buffer containing result information generated by speculative instructions;

a reservation station coupled to said reorder buffer, said reservation station containing information associated with instructions pending execution; and

an allocator coupled to said reorder buffer and said reservation station, said allocator allocating entries of said reorder buffer and said reservation station during an allocation pipestage of said pipelined processor, said allocation pipestage following a decode pipestage of said pipelined processor and preceding a dispatch pipestage of said pipelined processor, said allocator allocating entries of said reservation station for each pending instruction except for one or more register exchange instructions.

10. The computer system of claim 9 wherein the second processor further comprises a store buffer for containing information associated with store instructions, wherein entries of said store buffer are allocated during said allocation pipestage and remain allocated until said store instructions are executed.

11. The computer system of claim 10 wherein said allocator allocates entries of said store buffer.

12. The computer system of claim 11 wherein said allocator comprises:

sequential allocation logic to sequentially allocate vacant entries of said store buffer to instructions provided by said issue unit; and

nonsequential allocation logic to nonsequentially allocate entries of said reservation station to instructions provided by said issue unit.

13. The computer system of claim 10 further comprising:

a load buffer comprising a plurality of entries for containing information associated with load instructions, wherein entries of said load buffer are allocated during said allocation pipestage and remain allocated until said retirement pipestage; and

wherein said reorder buffer, said load buffer and said store buffer each comprise circular buffers having a tail pointer indicating a last valid entry and a head pointer indicating a first entry.

14. The computer system of claim 9 wherein said reorder buffer comprises a plurality of physical registers for containing said operand information.

15. In a pipeline processor having a pipeline comprising allocation, dispatch and retirement pipestages, a method of dynamically allocating resources in said processor, said method comprising the steps of:

(a) allocating entries of a first buffer to instructions during said allocation pipestage and until said retirement pipestage for said instructions, wherein said first buffer contains information associated with speculative instructions;

(b) allocating entries of a reservation station to instructions during said allocation pipestage and until said dispatch pipestage, wherein said reservation station contains information associated with instructions pending execution; and

wherein said step (b) of allocating entries of a reservation station is performed for each instruction processed by said processor except for one or more register exchange instructions.

16. The method of claim 15 further comprising the steps of:

(c) allocating entries of a load buffer to load instructions during said allocation pipestage and until said retirement pipestage;

(d) allocating entries of a store buffer to store instructions during said allocation pipestage and until said store instructions are performed over a bus coupled to said processor.

17. The method of claim 15 further comprising the steps of:

(c) allocating entries of a load buffer to load instructions during said allocation pipestage and until said retirement pipestage;

(d) allocating entries of a store buffer to store instructions during said allocation pipestage and until a time period subsequent to said retirement pipestage.

18. The method of claim 15 wherein step (a) is performed sequentially and step (b) is performed nonsequentially.

19. The method of claim 15 wherein step (a) is performed for each instruction processed by said processor.

20. The method of claim 19 wherein the one or more register exchange instructions include a floating point register exchange instruction.

21. The method of claim 15 further comprising the step of stalling said processor from issuing instructions if entries within said first buffer or said load buffer are unavailable.

22. The method of claim 16 further comprising the step of stalling said processor from issuing instructions if entries within said first buffer, reservation station, load buffer or said store buffer are unavailable.