16.2.4. The Three Stages of Control Transfer Instructions

Control transfer instructions can be divided into unconditional and conditional types. An unconditional jump directly moves the control to a specified memory location, whereas a conditional jump checks certain flags before proceeding.

Control transfer instructions are essential parts of programming that dictate where the flow of execution should go next. There are two primary types: unconditional and conditional. Unconditional jumps tell the CPU to move to a specified memory location without any conditions. For example, if a program executes a sequence of instructions and encounters an unconditional jump, it will jump to a predetermined section of code every time, regardless of any flags or conditions. In contrast, conditional jumps check specific conditions (like whether a variable equals zero) before moving to the next instruction. If the condition is met, the jump occurs; if not, the next sequential instruction is executed.

Think of control transfer instructions as traffic lights at intersections. An unconditional jump is like a green light, allowing cars to move freely without stopping. A conditional jump is like a red light, where the cars (or instructions) must stop and only move if the light changes (the conditions are met).

The first step involves loading the program counter (PC) into the memory address register to fetch the corresponding instruction from memory.

In the first stage of control transfer instructions, the CPU needs to know what instruction it should execute next, which is indicated by the program counter (PC). This stage begins by loading the value of the PC into the memory address register (MAR). Once this is accomplished, the CPU signals the memory to read that specific address. The instruction at that address, which could be a command to jump somewhere else, is then retrieved and placed into the memory data register (MDR). Crucially, the incrementing of the PC is paused because, in case of a jump, the CPU will not proceed to the next sequential instruction immediately. Instead, the current value of the PC is stored in a temporary register to allow for the potential control transfer to a different instruction.

Imagine you’re preparing a travel itinerary based on your current location. You note down a specific place (the address in the MAR) and check your map (the memory) for directions (instructions). If you see that you need to jump to a different location instead of moving to the next address, you momentarily pause to write down your current address in a notepad (temporary register), ensuring you can return to it later if needed.

In this stage, the updated address for the next instruction is calculated and held until the memory operation completes.

Once the instruction has been fetched, the second stage is all about preparing for the control transfer. The newly calculated address for the instruction that is to be executed next is derived from the information stored in the instruction register (IR). If the instruction is a jump, the offset for the jump is acquired from the IR and added to the current PC value stored in the temporary register. This new address, where the jump takes you, is held ready to be loaded into the PC once the memory operation is verified as successful. The value at this point is crucial as it dictates where the execution will proceed next.

Returning to our travel analogy, this stage is like evaluating your map after you’ve found your location (fetched the instruction). You now calculate how far off the next destination (the jump target) is based on the directions received. You keep this information handy until you confirm that it’s time to move on.

The final stage involves executing the instruction at the new given address by moving the PC to the target address.

In the last stage, the CPU executes the instruction that has been identified by the new address. The calculated address that was held in the temporary register is now loaded into the program counter (PC). With this complete, the CPU effectively jumps to the new instruction location. This stage is the transition from preparation to action where the program can now continue its execution based on the control transfer instruction that was just processed.

Completing the travel analogy, this final stage is like actually driving to the destination you’ve calculated. You’ve confirmed the directions, made necessary adjustments for any detours, and now you’re heading towards the new location indicated on your map (the instruction address).

--