Method to partition large code across multiple e-caches

1. A method for executing an instruction stream, comprising:

partitioning the instruction stream using a partition point to obtain a first partition of the instruction stream and a second partition of the instruction stream;

configuring the first partition and the second partition to fit separate e-caches;

analyzing the first partition and the second partition to generate a runtime execution facility and a data transfer facility; and

executing the first partition and the second partition on separate processors using the runtime execution facility and the data transfer facility.

2. The method of claim 1 , wherein the second partition executes after the first partition.

3. The method of claim 2 , wherein the second partition executes simultaneously with the first partition.

4. The method of claim 2 , further comprising:

storing the first partition and the second partition at separate addresses in a memory.

5. The method of claim 4 , further comprising:

retrieving the first partition and the second partition from the memory;

storing the first partition and the second partition in the separate e-caches; and

aligning a data segment in the memory.

6. The method of claim 5 , further comprising:

retrieving the first partition and the second partition from the separate e-caches.

7. The method of claim 1 , further comprising:

generating a plurality of basic blocks by analyzing the instruction stream, wherein the plurality of basic blocks is used in partitioning the instruction stream; and

assigning the first partition to a first processor and assigning the second partition to a second processor.

8. The method of claim 7 , wherein partitioning the instruction stream comprises:

locating a stride block within the instruction stream; and

locating the partition point within the stride block.

9. The method of claim 8 , wherein the partition point comprises an instruction of an unconditional basic block.

10. The method of claim 1 , wherein generating the runtime execution facility comprises:

determining an execution order of the first partition and the second partition;

generating an instruction to enforce execution order of the first partition and the second partition; and

inserting the instruction into the second partition.

11. The method of claim 10 , wherein the instruction implements a semaphore.

12. The method of claim 1 , wherein generating the data transfer facility comprises:

adding a STORE instruction to the first partition and adding a LOAD instruction to the second partition; and

allocating memory for a partition spill used for executing the first partition and the second partition.

13. The method of claim 12 , wherein generating the data transfer facility comprises:

executing the STORE instruction and the LOAD instruction to make data available for execution of the second partition.

14. The method of claim 1 , wherein a size of the instruction stream is greater than a size of the e-cache.

15. A method for executing an instruction stream, comprising:

partitioning the instruction stream using a partition point to obtain a first partition of the instruction stream and a second partition of the instruction stream;

configuring the first partition and the second partition to fit separate e-caches;

analyzing the first partition and the second partition to generate a runtime execution facility and a data transfer facility;

executing the first partition and the second partition on separate processors using the runtime execution facility and the data transfer facility;

retrieving the first partition and the second partition from the memory;

storing the first partition and the second partition in the separate e-caches;

aligning a data segment in the memory; and

retrieving the first partition and the second partition from the separate e-caches.

16. A system for executing an instruction stream, comprising:

a partitioning program configured to partition the instruction stream into a first partition of the instruction stream and a second partition of the instruction stream, store the first partition at a first address in a memory, and store the second partition at a second address in the memory;

wherein the first partition and the second partition are fitted to separate e-caches by the partitioning program;

a runtime execution facility configured to enforce execution order of the first partition and the second partition; and

a data transfer facility configured to transfer data required by execution of the second partition and supplied by execution of the first partition.

17. The system of claim 16 , further comprising:

a data segment in memory aligned with a third address in the memory.

18. The system of claim 16 , wherein the runtime execution facility comprises a semaphore.

19. The system of claim 16 , wherein the data transfer facility comprises a STORE instruction, a LOAD instruction, and a partition spill.

20. The system of claim 16 , wherein a size of a first e-cache of the separate e-caches is greater than a size of the first partition.

21. The system of claim 16 , further comprising:

a third partition of the instruction stream fitted to the first e-cache.

22. The system of claim 16 , wherein the first partition is included in a stride block.

23. A system for executing an instruction stream, comprising:

a partitioning program configured to partition the instruction stream into a first partition of the instruction stream and a second partition of the instruction stream, store the first partition at a first address in a memory, and store the second partition at a second address in the memory;

wherein the first partition and the second partition are fitted to separate e-caches by the partitioning program;

a runtime execution facility configured to enforce execution order of the first partition and the second partition;

a data transfer facility configured to transfer data required by execution of the second partition and supplied by execution of the first partition; and

a data segment in memory aligned with a third address in the memory.

24. An apparatus for executing an instruction stream, comprising:

means for partitioning the instruction stream using a partition point to obtain a first partition of the instruction stream and a second partition of the instruction stream;

means for configuring the first partition and the second partition to fit separate e-caches;

means for analyzing the first partition and the second partition to generate a runtime execution facility and a data transfer facility; and

means for executing the first partition and the second partition on separate processors using the runtime execution facility and the data transfer facility.