Merge Logic for Clustered Multithreaded VLIW Processors

Clustered VLIW embedded processors have become widespread due to benefits of simple hardware and low power. Simultaneous MultiThreading (SMT) is a well known technique that uses thread level parallelism at the instruction grain level. However, implementing SMT for VLIW requires complex structures. CSMT (cluster-level simultaneous MultiThreading) allows some degree of SMT in clustered VLIW processors with minimal hardware cost and complexity. This paper deals with the hardware required to implement CSMT instruction merge logic on a clustered VLIW processor. The paper presents two implementations of CSMT merge logic and an analysis of both comparing design issues like delay and number of transistors required.