Optimal deallocation of instructions from a unified pick queue

A method comprising:

• maintaining an array comprising a plurality of entries, wherein one or more entries may be allocated for use by any of a plurality of threads each clock cycle;
• determining a consumer instruction is dependent on a result of a producer instruction, wherein the producer instruction is allocated in a first entry of the plurality of entries and the consumer instruction is allocated in a second entry of the plurality of entries;
• determining the producer instruction has a fixed execution latency;
• picking the producer instruction for issue to execution resources in a first clock cycle;
• measuring an amount of time elapsed since the first clock cycle;
• and
• picking the consumer instruction for issue to execution resources in a second clock cycle based at least in part on determining an amount of time elapsed between the first clock cycle and the second clock cycle equals the fixed execution latency, wherein the second clock cycle is prior to an indication that the producer instruction has completed.

A microprocessor comprising:

• an instruction fetch unit configured to fetch instructions corresponding to a plurality of threads;
• a rename unit configured to rename the fetched instructions after being decoded;
• and
• a unified pick queue coupled to receive the renamed instructions;
• wherein the unified pick queue is further configured to:
• maintain an array comprising a plurality of entries, wherein one or more entries may be allocated for use by any of a plurality of threads each clock cycle;
• determine a consumer instruction is dependent on a result of a producer instruction, wherein the producer instruction is allocated in a first entry of the plurality of entries and the consumer instruction is allocated in a second entry of the plurality of entries;
• determine the producer instruction has a fixed execution latency;
• pick the producer instruction for issue to execution resources in a first clock cycle;
• measure an amount of time elapsed since the first clock cycle;
• and
• pick the consumer instruction for issue to execution resources in a second clock cycle based at least in part on determining an amount of time elapsed between the first clock cycle and the second clock cycle equals the fixed execution latency, wherein the second clock cycle is prior to an indication that the producer instruction has completed.