Multi-version register file for multithreading processors with live-in precomputation
Disclosed are selected embodiments of a processor that may include a plurality of thread units and a register file architecture to support speculative multithreading. For at least one embodiment, live-in values for a speculative thread are computed via execution of a precomputation slice and are stored in a validation buffer for later validation. A global register file holds the committed architecture state generated by a non-speculative thread. Each thread unit includes a local register file. A directory indicates, for each architectural register, which speculative thread(s) have generated a value for the architectural register. Other embodiments are also described and claimed.
1. A system, comprising:
a memory subsystem coupled to a processor;
the processor having N thread units to support speculative multithreading; and
a register versioning table coupled to the thread units, the register versioning table having N×R entries, the register versioning table to indicate, for each of the N thread units, whether the thread unit has written a value for each of R architectural registers;
wherein each thread unit includes a validation buffer to store speculatively precomputed live-in values;
wherein each thread unit further includes a local register file for the R architectural registers, wherein each thread unit determines whether the precomputed live-in values match local register file values, if such values match, the precomputed live-in values are marked valid; and
a global register file to hold the committed state for each of the R architectural registers, wherein the global register file is accessed by a thread unit when the thread unit performs a read to a register not available in its local register file and no less-speculative thread has written the register without committing it.
2. The system of claim 1 , wherein: the processor is a chip multiprocessor having a plurality of processing cores in a single chip package.
3. The system of claim 2 , wherein: each of the thread units is associated with a different of one of said processing cores.
4. The system of claim 1 , wherein: the processor is a simultaneous multithreading processor having a plurality of logical processors.
5. The system of claim 4 , wherein: each of the thread units is associated with a different of one of said logical processors.
6. The system of claim 1 , wherein: the memory subsystem includes a DRAM memory.
7. The system of claim 1 , wherein: the global register file is accessible to each of the thread units.