Granted Patent B1

US 7,406,587 · App. 10/210,257 · Granted Jul 29, 2008

Abstract

Claims (62)

1. A method for renaming registers in a microprocessor, comprising:

receiving an instruction for execution;

decoding the instruction;

placing the instruction in an active list;

mapping a logical register associated with the instruction to a selected unique physical register;

executing the instruction to obtain a result;

placing the result in the selected unique physical register;

moving the result to a floating point status register prior to graduation of the instruction;

returning an identity of the selected unique physical register after moving the result to the floating point status register for future selection upon graduation of the instruction in order to maintain a fixed order for selection of unique physical registers;

providing an indication in the mapping that the result is in the selected unique physical register while the instruction is still valid or in the floating point status register after the instruction graduates.

2. The method of claim 1 , further comprising:

retrieving the result from the floating point status register in response to an exception.

3. The method of claim 1 , further comprising:

receiving a next instruction for execution;

decoding the next instruction;

mapping a logical register associated with the next instruction to a next unique physical register in consecutive succession with the selected unique physical register;

executing the next instruction to obtain a next result;

placing the next result in the next unique physical register;

moving the next result to the floating point status register prior to graduation of the next instruction;

returning an identity of the next unique physical register for future selection upon graduation of the next instruction.

4. The method of claim 1 , further comprising:

providing an indication that the selected unique physical register is not available for use.

5. The method of claim 3 , further comprising:

providing an indication that the next result is in the next unique physical register while the next instruction is still valid or in the floating point status register after the instruction graduates.

6. The method of claim 5 , wherein the indication is maintained in a single bit status register.

7. The method of claim 1 , further comprising:

storing the mapping of the logical register to the selected unique physical register and the indication as to where the result is stored in response to the instruction being a conditional branch instruction.

8. The method of claim 1 , further comprising:

obtaining an identity of the selected unique physical register from a list of available unique physical register identities.

9. The method of claim 1 , further comprising:

placing condition codes associated with the instruction in the active list;

using a pointer corresponding to the active list to identify the mapping of the logical register to the selected unique physical register.

10. The method of claim 1 , wherein mapping of the logical register to the selected unique physical register is performed without any prior mapping initialization.

11. A system for renaming registers, comprising:

a plurality of unique physical registers;

a mapping table operable to map a logical register associated with an instruction to a selected one of the plurality of unique physical registers, the selected unique physical register operable to receive a result in response to execution of the instruction;

an active list operable to maintain condition codes associated with the instruction;

a floating point status register operable to receive the result from the selected unique physical register prior to graduation of the instruction;

wherein the mapping table includes an indication that the result is in the floating point status register after the instruction graduates or in the selected unique physical register while the instruction is still valid, the mapping table operable to return an identity of the selected one of the plurality of unique physical registers after moving the result to the floating point status register for future selection upon graduation of the instruction in order to maintain a fixed order for selection of unique physical registers.

12. The system of claim 11 , wherein the mapping table and the active list are accessed through a same write pointer.

13. The system of claim 11 , wherein the selected unique physical register is returned for future reuse.

14. The system of claim 13 , wherein the plurality of unique physical registers are selected and returned in a consecutive order.

15. The system of claim 11 , further comprising:

a shadow mapping table operable to receive the mapping of the logical register to the selected unique physical register and an indication as to whether the result is in the floating point status register or the selected unique physical register.

16. A system for renaming registers, comprising:

means for receiving an instruction for execution;

means for decoding the instruction;

means for placing the instruction in an active list;

means for mapping a logical register associated with the instruction to a selected unique physical register;

means for executing the instruction to obtain a result;

means for placing the result in the selected unique physical register;

means for moving the result to a floating point status register prior to graduation of the instruction;

means for returning an identity of the selected unique physical register after moving the result to the floating point status register for future selection upon graduation of the instruction in order to maintain a fixed order for selection of unique physical registers;

means for providing an indication in the mapping that the result is in the selected unique physical register while the instruction is still valid or in the floating point status register after the instruction graduates.

17. The system of claim 16 , further comprising:

means for storing the mapping of the logical register to the selected unique physical register and the indication as to where the result is stored in response to the instruction being a conditional branch instruction.

18. The system of claim 16 , further comprising:

means for providing an indication that the result is either in the selected unique physical register or in the floating point status register.

19. The system of claim 18 , wherein the indication is maintained in a single bit status register.

20. The system of claim 16 , further comprising:

means for placing condition codes associated with the instruction in the active list;

means for using a pointer corresponding to the active list to identify the mapping of the logical register to the selected unique physical register.

Assignments (13)