Why microprogrammed control is more flexible

A microprogrammed control unit is more flexible because its behaviour is defined by a microprogram stored in the control memory, not by fixed wiring.

• To change or add an instruction, only the contents of the control memory are rewritten; no hardware is rebuilt.
• Errors can be corrected by changing the microcode.
• With a writable control store, the same hardware can emulate a different instruction set.
• New instructions can easily be added to a family of machines, and the design is systematic.

In a hardwired unit, the control signals come from gates and flip-flops, so any change needs a redesign of the logic and new hardware.

Sequencing technique used in control memory

The sequencer selects the address of the next microinstruction from:

1. CAR + 1 (normal sequential execution),
2. Branch address (AD) for conditional or unconditional jumps (the CD field tests a status bit and the BR field selects JMP),
3. Mapped address from the instruction opcode (0 xxxx 00) at the end of the fetch routine,
4. Return address from the subroutine register (SBR) on RET, saved there by CALL.

 opcode -> map ----+
 AD field ---------+--> MUX --> CAR --> Control memory
 SBR --------------+      ^                 |
 CAR+1 ----------+--------+                 v
                 ^                      Control data
 status, CD, BR -+ (branch logic)        register

CALL stores CAR + 1 in SBR, and RET returns it to CAR. Initially, CAR is loaded with the address of the fetch routine.