PROCESSOR INCLUDING HYBRID REDUNDANCY FOR LOGIC ERROR PROTECTION

• US 20090183035A1
• Filed: 01/10/2008
• Published: 07/16/2009
• Est. Priority Date: 01/10/2008
• Status: Abandoned Application

• Current Assignee: GlobalFoundries, Inc.
• Original Assignee: GlobalFoundries, Inc.
• Inventors: Quach, Nhon, Butler, Michael G.
• Application Number:
  US11/972,166
• Publication Number:
  US 20090183035A1
• Time in Patent Office: Days
• US Class Current: 714/48
• CPC Class Codes: G06F 11/1405 at machine instruction level
• CPC Class Codes: G06F 11/1497 Details of time redundant execution on a single processing unit
• CPC Class Codes: G06F 11/1641 where the comparison is not performed by the redundant processing components
• CPC Class Codes: G06F 11/1645 and the comparison itself uses redundant hardware

Abstract

A processor core includes an instruction decode unit that may dispatch a same integer instruction stream to a plurality of integer execution units and may consecutively dispatch a same floating-point instruction stream to a floating-point unit. The integer execution units may operate in lock-step such that during each clock cycle, each respective integer execution unit executes the same integer instruction. The floating-point unit may execute the same floating-point instruction stream twice. Prior to the integer instructions retiring, compare logic may detect a mismatch between execution results from each of the integer execution units. In addition, prior to the results of the floating-point instruction stream transferring out of the floating-point unit, the compare logic may also detect a mismatch between results of execution of each consecutive floating-point instruction stream. Further, in response to detecting any mismatch, the compare logic may cause instructions causing the mismatch to be re-executed.

Claims

1. A processor core configured to operate in a reliable execution mode, the processor core comprising:

an instruction decode unit configured to dispatch a same integer instruction stream to a plurality of integer execution units and to consecutively dispatch a same floating-point instruction thread to a floating point unit;

wherein the plurality of integer execution units is configured to operate in lock-step such that during each clock cycle, the plurality of integer execution units executes a same integer instruction;

wherein the floating-point unit is configured to execute the same floating-point instruction stream twice;

compare logic coupled to the plurality of integer execution units and the floating-point unit, wherein prior to instructions in the same integer instruction stream retiring, the compare logic is configured to detect a mismatch between execution results from each of the plurality of integer execution units;

wherein prior to the floating-point unit transferring the execution results of the floating-point instruction stream out of the floating-point unit, the compare logic is further configured to detect a mismatch between results of execution of each consecutive floating-point instruction stream;

wherein in response to the compare logic detecting any mismatch, the compare logic is configured to cause instructions causing the mismatch to be re-executed.

2. The processor core as recited in claim 1, wherein the plurality of integer execution units comprises a plurality of integer execution clusters, each including one or more first integer execution units and one or more first scheduler units.

3. The processor core as recited in claim 2, wherein the compare logic is further configured to compare signals corresponding to execution results of a first execution cluster of the plurality of integer execution clusters against signals corresponding to execution results of a second execution cluster of the plurality of integer execution clusters.

4. The processor core as recited in claim 3, wherein the compare logic comprises a distributed compare function included within the first execution cluster, the second execution cluster, and the floating point unit.

5. The processor core as recited in claim 1, further comprises a retirement queue coupled to the plurality of integer execution units and to the floating-point unit, wherein the retirement queue includes error protection logic configured to detect errors using an error detecting code.

6. The processor core as recited in claim 1, wherein the signals corresponding to the execution results execution results from each of the plurality of integer execution units comprise signatures generated from result signals conveyed on result buses within each of the plurality of integer execution units.

7. The processor core as recited in claim 1, wherein the compare logic is further configured to check results of branch mis-predictions within each of the plurality of integer execution units.

8. The processor core as recited in claim 1, wherein the compare logic is further configured to check for substantially simultaneous accesses to a level two cache by each of the plurality of integer execution units.

9. The processor core as recited in claim 1, wherein the instruction decode unit, the plurality of integer execution units, and the floating-point unit are configured to operate in the reliable execution mode dependent upon a voltage reference applied to an external device pin.

10. A method for protecting against logic errors in a processor core, the method comprising:

operating the processor core in a reliable execution mode;

dispatching a same integer instruction stream to a plurality of integer execution units and consecutively dispatching a same floating-point instruction stream to a floating point unit;

operating the plurality of integer execution units in lock-step such that during each clock cycle, the plurality of integer execution units executes a same integer instruction;

the floating-point unit executing the same floating-point instruction stream twice;

compare logic performing a comparison and detecting a mismatch between execution results from each of the plurality of integer execution units prior to instructions in the same integer instruction stream retiring;

compare logic performing a comparison and detecting a mismatch between results of execution of each consecutive floating-point instruction stream prior to the floating-point unit transferring the execution results of the floating-point instruction stream out of the floating-point unit;

re-executing instructions caused the mismatch in response to detecting any mismatch.

11. The method as recited in claim 10, wherein the plurality of integer execution units comprises a plurality of integer execution clusters, each including one or more first integer execution units and one or more first scheduler units.

12. The method as recited in claim 11, further comprising the compare logic comparing signals corresponding to execution results of a first execution cluster of the plurality of integer execution clusters against signals corresponding to execution results of a second execution cluster of the plurality of integer execution clusters.

13. The method e as recited in claim 12, wherein the compare logic comprises a distributed compare function included within the first execution cluster, the second execution cluster, and the floating point unit.

14. The method as recited in claim 10, further comprising detecting errors within a retirement queue using error detecting logic and an error detecting code.

15. The method as recited in claim 10, further comprising generating the signals corresponding to the execution results execution results from each of the plurality of integer execution units by generating signatures from result signals conveyed on result buses within each of the plurality of integer execution units.

16. The method as recited in claim 10, further comprising the compare logic checking results of branch mis-predictions within each of the plurality of integer execution units and providing an indication in response to detecting a branch misprediction on one integer execution unit and not another integer execution unit.

17. The method as recited in claim 10, further comprising the compare logic checking for substantially simultaneous accesses to a level two cache by each of the plurality of integer execution units and providing an indication in response to non-substantially simultaneous accesses.

18. The method as recited in claim 10, wherein the instruction decode unit, the plurality of integer execution units, and the floating-point unit operating in the reliable execution mode dependent upon a voltage reference applied to an external device pin.

19. A processor comprising:

a plurality of processor cores, each processor core comprising;

an instruction decode unit configured to dispatch a same integer instruction stream to a plurality of integer execution units and to consecutively dispatch a same floating-point instruction thread to a floating point unit;

wherein the plurality of integer execution units is configured to operate in lock-step such that during each clock cycle, the plurality of integer execution units executes a same integer instruction;

wherein the floating-point unit is configured to execute the same floating-point instruction stream twice;

compare logic coupled to the plurality of integer execution units and the floating-point unit, wherein prior to instructions in the same integer instruction stream retiring, the compare logic is configured to detect a mismatch between execution results from each of the plurality of integer execution units;

wherein prior to the floating-point unit transferring the execution results of the floating-point instruction stream out of the floating-point unit, the compare logic is further configured to detect a mismatch between results of execution of each consecutive floating-point instruction stream;

wherein in response to the compare logic detecting any mismatch, the compare logic is configured to cause instructions causing the mismatch to be re-executed.

20. The processor as recited in claim 19, wherein the plurality of integer execution units comprises a plurality of integer execution clusters, and wherein the compare logic is further configured to compare signals corresponding to execution results of a first execution cluster of the plurality of integer execution clusters against signals corresponding to execution results of a second execution cluster of the plurality of integer execution clusters.