⭐ Symmetric Multiprocessing (SMP)

1. Definition

Symmetric Multiprocessing (SMP) is a computer architecture in which two or more identical processors are connected to a single shared main memory and operate under a single operating system, sharing I/O and memory resources equally.

In SMP, all processors are peers—any processor can execute any task, and the OS can schedule tasks on any processor.

---

2. Key Features of SMP

1. Multiple processors: Two or more identical CPUs.
2. Shared memory: All processors have equal access to the main memory.
3. Single OS instance: One operating system controls all processors.
4. Uniform access time: Memory access time is the same for all processors (unlike NUMA).
5. Common I/O subsystem: All processors share the same I/O devices.

---

3. Benefits of SMP

1. Improved performance and throughput: Multiple processors can execute tasks concurrently.
2. Scalability: Adding more processors increases system performance (up to limits).
3. Reliability and fault tolerance: If one processor fails, others can continue processing (though not fully fault-tolerant).
4. Simpler programming model: Single memory space makes it easier to write concurrent programs.

---

4. Challenges of SMP

1. Cache coherency: Each processor typically has its own cache, so mechanisms (like MESI protocol) are needed to maintain consistent data.
2. Contention for memory and I/O: Multiple processors accessing the same memory or I/O can cause bottlenecks.
3. Synchronization overhead: Threads/processes must be synchronized to avoid race conditions.
4. Scalability limits: Performance gain diminishes as the number of processors increases due to contention.

---

5. Typical SMP Architecture

• Processors (P1, P2, …): Identical CPUs, each with private registers and possibly private cache.
• Shared Main Memory: Single memory accessible by all CPUs.
• System Bus / Interconnect: Connects processors to memory and I/O.
• Operating System: Single OS schedules tasks across all processors.

---

6. Comparison with Other Architectures

Feature	SMP	NUMA (Non-Uniform Memory Access)
Memory Access	Uniform (same latency for all)	Non-uniform (different latency)
Processors	Symmetric, peer CPUs	Nodes with local memory
Scalability	Limited due to bus contention	Better for large-scale systems
Programming Model	Simpler	More complex (awareness of memory locality needed)

---

7. Example Scenario

• Suppose an SMP system has 4 CPUs and shared memory.
• OS can schedule Thread 1 on CPU1, Thread 2 on CPU2, and so on.
• All threads access shared variables in memory.
• Cache coherency protocols ensure all CPUs see the most recent value of shared variables.

---

8. Exam-Friendly Summary

• SMP: Multiple identical processors sharing memory and I/O under a single OS.
• Goal: Increase throughput and parallelism for multiple tasks/threads.
• Key challenges: Cache coherency, memory contention, synchronization.
• Advantages: Simplicity, performance, scalability (moderate), reliability.