π Input-Output Interface
What is I/O Interface?
The I/O interface provides a method for transferring information between internal storage (memory/CPU) and external I/O devices. It resolves the differences between the computer and peripheral devices.
π‘ Key Concept: I/O interface acts as a bridge between the CPU and peripheral devices, handling timing, data format, and control signal differences.
Functions of I/O Interface
1. Data Buffer π¦
Temporarily stores data being transferred between the CPU and peripherals to compensate for speed differences.
2. Error Detection π
Detects and reports errors during data transmission using parity bits, checksums, or CRC.
3. Signal Conversion π
Converts signals between digital (CPU) and analog (some peripherals) formats when necessary.
4. Control & Timing β±οΈ
Coordinates timing and control signals between CPU and peripherals with different operating speeds.
I/O Interface Architecture
CPU
β
I/O Interface
β
Peripheral
Data flows through the interface which handles all communication complexities
Components of I/O Interface
- Data Register - Holds data being transferred
- Status Register - Contains status information about the device
- Control Register - Receives control commands from CPU
- Address Decoder - Selects the appropriate device
π Asynchronous Data Transfer
Understanding Asynchronous Transfer
Asynchronous data transfer allows two independent units to communicate without using a common clock. Each unit operates at its own speed, and synchronization is achieved through control signals.
Methods of Asynchronous Transfer
1. Strobe Control
Uses a single control line (strobe signal) to indicate when data is valid.
Simple
One-way
Fast
Types:
- Source-Initiated: Source sends data and activates strobe
- Destination-Initiated: Destination requests data using strobe
2. Handshaking
Uses two control signals for reliable bidirectional communication between devices.
Reliable
Two-way
Flexible
Handshaking Sequence
Step 1: Source places data on data bus
Step 2: Source activates Data Valid signal
Step 3: Destination reads data
Step 4: Destination activates Data Accepted signal
Step 5: Source removes data and deactivates Data Valid
Step 6: Destination deactivates Data Accepted
| Feature |
Strobe Control |
Handshaking |
| Control Lines |
1 |
2 |
| Reliability |
Moderate |
High |
| Speed |
Faster |
Slower |
| Acknowledgment |
No |
Yes |
π‘ Why Asynchronous? Asynchronous transfer is essential when devices operate at different speeds or when timing cannot be precisely synchronized across the system.
π Modes of Data Transfer
1. Programmed I/O
The CPU is totally responsible for data transfer. It continuously checks the status of the I/O device.
Characteristics:
- CPU monitors device status continuously (polling)
- CPU transfers each data item
- Simple to implement but inefficient
- CPU time is wasted waiting for devices
Programmed I/O Flow
CPU
β
Check Status
β
Transfer Data
β
Repeat
2. Interrupt-Initiated I/O
The device notifies the CPU when it's ready via an interrupt signal. The CPU can perform other tasks until interrupted.
Characteristics:
- No need for continuous polling
- CPU can execute other programs
- More efficient than programmed I/O
- Requires interrupt handling mechanism
Efficient
Event-Driven
Multitasking
3. Direct Memory Access (DMA)
Transfers large blocks of data directly between memory and I/O device without CPU intervention.
Characteristics:
- Fastest data transfer method
- DMA controller manages the transfer
- CPU is free to perform other tasks
- Ideal for bulk data transfers
High Speed
Block Transfer
CPU Independent
| Mode |
CPU Involvement |
Speed |
Best For |
| Programmed I/O |
High (100%) |
Slow |
Simple devices, small data |
| Interrupt I/O |
Medium |
Moderate |
Interactive devices |
| DMA |
Low (minimal) |
Fast |
Bulk data transfer |
β‘ Priority Interrupt
What is Priority Interrupt?
A priority interrupt system handles multiple interrupt requests by assigning priorities to different devices. When multiple interrupts occur simultaneously, the system serves the highest priority request first.
π‘ Key Concept: Priority systems ensure critical devices get immediate attention while less critical devices wait their turn.
Priority Interrupt Methods
1. Polling (Software)
CPU checks each device in sequence according to priority order.
- Simple to implement
- Flexible priority assignment
- Slower response time
- CPU cycles wasted in checking
2. Daisy Chain (Hardware)
Devices are connected in series. Interrupt acknowledge signal passes through devices in priority order.
- Fast response
- Minimal hardware cost
- Fixed priority based on position
- Single interrupt line
Daisy Chain Configuration
CPU
β
Device 1 (Highest)
β
Device 2
β
Device 3
3. Parallel Priority (Hardware)
Each device has its own interrupt line. Priority logic determines which request to serve.
- Fastest method
- Individual interrupt lines
- Complex hardware
- Most flexible priority control
Fastest
Most Expensive
Interrupt Service Routine (ISR)
Steps in Interrupt Handling π§
- Save current program state (PC, registers)
- Identify interrupt source
- Jump to appropriate ISR
- Execute ISR (handle the interrupt)
- Restore program state
- Return to interrupted program
β οΈ Important: Nested interrupts occur when a higher priority interrupt interrupts the ISR of a lower priority interrupt.
πΎ Direct Memory Access (DMA)
What is DMA?
DMA is a method that allows I/O devices to transfer data directly to/from memory without CPU intervention. A DMA controller manages the entire transfer process.
DMA Transfer Architecture
I/O Device
β
DMA Controller
β
Memory
CPU only initializes the transfer; DMA handles the rest
DMA Controller Registers
- Address Register: Holds memory address for transfer
- Word Count Register: Number of words to transfer
- Control Register: Mode of transfer (read/write)
- Status Register: Indicates transfer status
DMA Transfer Modes
1. Burst Mode (Block Transfer)
DMA controller takes over system bus and transfers entire block of data in one burst.
- Fastest transfer method
- CPU is halted during transfer
- Ideal for high-speed devices
Fastest
CPU Blocked
2. Cycle Stealing Mode
DMA controller transfers one word at a time, stealing a bus cycle when needed.
- CPU can continue execution
- Slightly slower than burst mode
- Minimal CPU disruption
Balanced
CPU Friendly
3. Transparent Mode
DMA transfers only when CPU is not using the bus.
- No CPU cycle stealing
- Slowest DMA mode
- Zero impact on CPU
Slowest
Zero CPU Impact
DMA Transfer Steps
Click to see DMA process flow π
- Initialization: CPU programs DMA controller with address, count, and control
- Request: I/O device requests DMA transfer
- Grant: DMA controller requests bus from CPU
- Transfer: DMA manages data transfer directly
- Complete: DMA sends interrupt to CPU when done
π‘ Advantage: DMA frees the CPU to perform other tasks, resulting in significant performance improvement for bulk data transfers.
π₯οΈ Input-Output Processor (IOP)
What is an IOP?
An Input-Output Processor (IOP) is a specialized processor designed to handle all I/O operations independently. It has its own instruction set and can execute complex I/O programs.
π‘ Evolution: IOP represents the highest level of I/O sophistication: Programmed I/O β Interrupt I/O β DMA β IOP
IOP vs CPU
| Aspect |
CPU |
IOP |
| Primary Function |
General computation |
I/O operations only |
| Instructions |
General purpose |
I/O specific |
| Memory Access |
Full access |
Shared with CPU |
| Operation |
Independent |
Receives commands from CPU |
Types of IOP
1. Channel
A simplified IOP that can execute limited I/O programs stored in memory.
- Operates using channel programs
- Manages multiple devices
- Common in mainframe computers
2. I/O Module
More sophisticated processor with its own local memory and instruction set.
- Can execute complex I/O tasks
- Has private memory for I/O tables
- Reduces CPU workload significantly
CPU-IOP Communication
Communication Flow
CPU
β
Memory
β
IOP
β
I/O Devices
Communication Methods
1. Memory-Mapped Communication
- CPU and IOP communicate through shared memory
- CPU writes commands to memory locations
- IOP reads commands and writes status back
- Mailbox system for message passing
2. Interrupt Communication
- IOP interrupts CPU when task completes
- CPU can interrupt IOP with new commands
- Enables asynchronous operation
IOP Instruction Format
Typical IOP Instructions π
- Read/Write: Transfer data between device and memory
- Sense: Check device status
- Jump: Branch based on conditions
- Control: Send control signals to devices
- Transfer in Channel: Move data within channel
β‘ Performance Boost: IOPs can handle complex I/O operations including data formatting, error checking, and device management without CPU involvement.
π‘ Serial Communication
What is Serial Communication?
Serial communication transmits data one bit at a time over a single channel or wire. It's widely used for long-distance communication and connecting peripheral devices.
Serial vs Parallel Communication
| Feature |
Serial |
Parallel |
| Data Lines |
1 or 2 |
8, 16, 32+ |
| Speed (short distance) |
Slower |
Faster |
| Cost |
Low |
High |
| Distance |
Long |
Short |
| Crosstalk |
Minimal |
Significant |
| Use Case |
Networks, peripherals |
Internal buses |
Types of Serial Communication
1. Synchronous Serial Communication
Data is transmitted with a clock signal. Both sender and receiver share the same clock.
Characteristics:
- Requires clock synchronization
- Faster and more efficient
- Used for high-speed applications
- Examples: SPI, IΒ²C, SATA
Fast
Clock Required
High Efficiency
2. Asynchronous Serial Communication
Data is transmitted without a clock signal. Uses start and stop bits for synchronization.
Characteristics:
- No clock signal needed
- Uses start and stop bits
- Flexible but less efficient
- Examples: UART, RS-232, USB
Simple
No Clock
Flexible
Asynchronous Data Frame
Frame Structure:
[Start Bit (1)] [Data Bits (5-9)] [Parity Bit (0-1)] [Stop Bits (1-2)]
Serial Communication Parameters
1. Baud Rate β‘
Number of signal changes per second. Common rates: 9600, 19200, 38400, 57600, 115200 bps.
Note: Baud rate must match between sender and receiver!
2. Data Bits π
Number of bits in each data packet (typically 5, 6, 7, or 8 bits). Most common is 8 bits.
3. Parity Bit π
Used for error detection. Types:
- Even Parity: Makes total 1s even
- Odd Parity: Makes total 1s odd
- None: No parity checking
4. Stop Bits βΈοΈ
Indicates end of transmission (1, 1.5, or 2 bits). Provides time for receiver to prepare for next byte.
Serial Communication Standards
RS-232
- Old but still widely used standard
- Point-to-point communication
- Maximum distance: ~15 meters
- Voltage levels: Β±3V to Β±15V
USB (Universal Serial Bus)
- Modern, high-speed serial standard
- Plug-and-play capability
- Supports multiple devices (up to 127)
- Provides power to devices
- Versions: USB 1.1, 2.0, 3.0, 3.1, 4.0
UART (Universal Asynchronous Receiver-Transmitter)
- Hardware device for asynchronous serial communication
- Converts parallel data to serial and vice versa
- Widely used in embedded systems
- Full-duplex communication
π‘ Modern Trend: Despite being "slower," serial communication dominates modern computing due to lower cost, longer distance capability, and higher speeds at modern frequencies.