Programmed I/O vs Interrupt-Driven I/O: Differences Explained

Input/Output (I/O) operations are fundamental to computer systems. They allow the system to interact with external devices such as keyboards, mice, printers, and other peripherals. Two common methods used to perform I/O operations are programmed I/O and interrupt-driven I/O. In this article, we will explore the difference between these two methods and how they are implemented.

Programmed I/O:

Programmed I/O is a simple method of performing I/O operations. In programmed I/O, the processor directly communicates with the I/O device to transfer data. This method involves the following steps:

• The processor sends a request to the I/O device to transfer data.
• The I/O device responds to the request by sending or receiving data.
• The processor waits for the I/O operation to complete before moving on to the next instruction.

Programmed I/O is a synchronous operation, which means that the processor waits for the I/O device to respond before it can execute the next instruction. This method is suitable for simple I/O operations that involve small amounts of data.

Interrupt-driven I/O:

Interrupt-driven I/O is a more sophisticated method of performing I/O operations. In this method, the processor does not wait for the I/O device to respond. Instead, the I/O device sends an interrupt signal to the processor when it is ready to send or receive data. The processor then suspends its current task and services the interrupt request. Once the interrupt request is serviced, the processor resumes its previous task.

Interrupt-driven I/O is an asynchronous operation, which means that the processor does not have to wait for the I/O device to respond. This method is well suited to slow or unpredictable devices such as keyboards, mice and printers, because the CPU can do useful work while it waits. For large block transfers, such as from disk to memory, DMA is the better choice (see below).

Difference between Programmed I/O and Interrupt-driven I/O:

The main difference between programmed I/O and interrupt-driven I/O is that in programmed I/O, the processor waits for the I/O device to respond before moving on to the next instruction. In interrupt-driven I/O, the processor does not wait for the I/O device to respond. Instead, it services the interrupt request and then resumes its previous task.

Programmed I/O vs Interrupt-driven I/O: Comparison Table

Feature	Programmed I/O	Interrupt-driven I/O
What the CPU does while waiting	Busy-waits, repeatedly checking the device status flag (polling)	Runs other tasks until the device sends an interrupt
Who starts each transfer	The CPU, after it sees the device is ready	The device, by raising an interrupt request
CPU efficiency	Low: many cycles are wasted in the polling loop	High: no time is wasted waiting
Hardware needed	Simple, no interrupt logic	Interrupt lines and often an interrupt controller (e.g. 8259 PIC)
Overhead per transfer	Very small	Context save/restore and ISR call on every interrupt
Does data pass through the CPU?	Yes	Yes, every byte or word
Best for	Fast or simple devices, short transfers, small embedded systems	Slow or unpredictable devices (keyboard, mouse, printer)
Example	A loop that keeps reading a UART status bit until a byte arrives	A key press on a PC raising a hardware interrupt (INT 09h)

Advantages and Disadvantages

Programmed I/O

• Advantage: Very simple to design and program.
• Advantage: No extra interrupt hardware is required.
• Advantage: Predictable timing, which is useful in simple embedded systems.
• Disadvantage: Wastes CPU time while busy-waiting.
• Disadvantage: The CPU cannot do other work during the transfer.

Interrupt-driven I/O

• Advantage: CPU time is used efficiently; no busy-waiting.
• Advantage: Allows multitasking and quick response to external events.
• Disadvantage: Needs extra hardware and more complex software (ISRs, priorities).
• Disadvantage: Each interrupt adds overhead for saving and restoring the CPU state.
• Disadvantage: The CPU still moves every byte, so it is slow for large block transfers.

How an Interrupt Is Serviced (Step by Step)

1. The I/O device raises an interrupt request (IRQ) when it is ready.
2. The CPU finishes the instruction it is currently executing.
3. If interrupts are enabled (the IF flag in the 8086), the CPU acknowledges the request.
4. The CPU saves its context. In the 8086 it pushes the FLAGS, CS and IP registers onto the stack.
5. The CPU finds the address of the Interrupt Service Routine (ISR). In the 8086 this comes from the interrupt vector table at address type × 4.
6. The ISR runs and transfers the data to or from the device.
7. The ISR ends with IRET, which restores IP, CS and FLAGS, and the CPU resumes the interrupted program.

What About DMA (Direct Memory Access)?

Both programmed I/O and interrupt-driven I/O move every byte through the CPU. For large blocks of data, such as reading a disk sector, that is slow. With DMA, a DMA controller (for example the Intel 8237) transfers the whole block directly between the device and memory. The CPU only sets up the transfer at the start and receives one interrupt when it is finished.

Method	CPU involvement	Best for
Programmed I/O	During the whole transfer, including waiting	Simple, low-speed tasks
Interrupt-driven I/O	For every byte or word, but no waiting	Slow, unpredictable devices
DMA	Only at the start and the end	Large, high-speed block transfers

Practice MCQs for Job and University Exams

1. In programmed I/O, the CPU:
(a) Is interrupted by the device when it is ready
(b) Repeatedly checks the status of the device
(c) Is bypassed by a DMA controller
(d) Does not take part in data transfer
Answer: (b)

2. Which I/O method wastes the most CPU time?
(a) Programmed I/O
(b) Interrupt-driven I/O
(c) DMA
(d) All waste equal time
Answer: (a)

3. Which method lets the CPU run other instructions while waiting for a slow device?
(a) Programmed I/O
(b) Interrupt-driven I/O
(c) Polling
(d) Busy-waiting
Answer: (b)

4. The best method for transferring a large block of data from disk to memory is:
(a) Programmed I/O
(b) Interrupt-driven I/O
(c) DMA
(d) Polling
Answer: (c)

5. In the 8086, the ISR address for interrupt type n is stored at memory address:
(a) n × 2
(b) n × 4
(c) n × 8
(d) n + 4
Answer: (b)

6. Which 8086 instruction returns from an Interrupt Service Routine?
(a) RET
(b) IRET
(c) HLT
(d) INT
Answer: (b)

7. Repeatedly checking a device status flag in a loop is called:
(a) Polling
(b) Vectoring
(c) Masking
(d) Nesting
Answer: (a)

8. In interrupt-driven I/O, data between the I/O device and memory passes through:
(a) The DMA controller
(b) The CPU
(c) The cache only
(d) The interrupt controller
Answer: (b)

Frequently Asked Questions (FAQs)

Q: Is polling the same as programmed I/O?
A: Almost. Programmed I/O normally uses polling: the CPU keeps reading the device status register until the device is ready, then transfers the data.

Q: Which is faster, programmed I/O or interrupt-driven I/O?
A: For a single short transfer with a fast device, polling can respond slightly faster because there is no context-saving overhead. For the system as a whole, interrupt-driven I/O is better because the CPU is not stuck waiting.

Q: Why do we need DMA if interrupts already exist?
A: With interrupts, the CPU still moves every byte and handles an interrupt each time. For large blocks this overhead is too high, so a DMA controller moves the data directly to memory.

Q: What is an Interrupt Service Routine (ISR)?
A: An ISR is a short program that runs when an interrupt occurs. It handles the device (for example, reads the received byte) and then returns control to the interrupted program.

Q: What is interrupt latency?
A: Interrupt latency is the time between a device raising an interrupt and the CPU starting to execute its ISR. Lower latency means faster response to external events.

📚 Keep Learning

• Importance of I/O Module
• Features of the Intel 8086 Microprocessor: A Beginner's Guide
• Assembly Language Programming Codes for the 8086 Microprocessor
• 8086 Assembly Tutorial: Run Your First MASM Program in DOSBox

Conclusion

In conclusion, both programmed I/O and interrupt-driven I/O are methods used to perform I/O operations. Programmed I/O is a simple method suitable for simple I/O operations that involve small amounts of data. Interrupt-driven I/O is more efficient for slow or unpredictable devices, because the CPU only steps in when the device is ready. By understanding the difference between these two methods, developers can choose the appropriate method for their application to optimize system performance.

Related Posts

• What Is an I/O Module and Why Is It Needed?
• Features of Intel 8086 Microprocessor: Architecture Basics for Beginners
• What Is a Computer Interface? Hardware vs Software Interface
• Programming Language Acronyms and Full Forms (C, PHP, AJAX, COBOL & More)