Text this: Optimizing Instruction Scheduling and Register Allocation for Register-File-Connected Clustered VLIW Architectures