Software controlled pre-execution in a multithreaded processor

A processor, comprising:

• a fetch unit capable of fetching instructions from a plurality of threads;
• a program counter coupled to said fetch unit;
• an instruction cache coupled to said fetch unit;
• a data cache coupled to said instruction cache;
• wherein said processor processes a program in a first thread that includes instructions which cause the processor to spawn a second thread in which at least a portion of the same program also executes, said second thread runs concurrently with the first thread, but ahead of the first thread in program order.

A method of running a program in a processor, comprising:

• (a) inserting pre-execution thread instructions in the program;
• (b) spawning a pre-execution thread when designated by the inserted instructions;
• and
• (c) running said pre-execution thread concurrently with a main thread wherein both the pre-execution and the main threads include instructions from the same program, the pre-execution thread running ahead of the main thread.

A processor, comprising:

• a fetch unit capable of fetching instructions from a plurality of threads;
• a program counter coupled to said fetch unit;
• an instruction cache coupled to said fetch unit;
• a data cache coupled to said instruction cache;
• wherein said processor pre-executes instructions from a main thread.