Study sheet: CPU Performance Factors

📋 Course Outline

1. Clock Speed
2. CPU Cores
3. Cache Memory
4. Parallel Processing
5. Threading Models
6. Combined Performance and Bottlenecks

📖 1. Clock Speed

🔑 Key Concepts & Definitions

• Clock speed : the frequency at which the CPU operates

★ Must-know

📌 A higher clock speed generally allows a CPU to process more instructions per second and complete instructions faster when otherwise identical CPUs are compared.

Further detail

📌 Increasing clock speed also increases power consumption and heat production.

• Clock speed is measured in Hertz (Hz), and 1 Hz equals 1 clock cycle per second.

💡 Memory Hook

Higher frequency → more instructions per second → faster processing, but also more heat and power use.

📖 2. CPU Cores

🔑 Key Concepts & Definitions

• CPU core : an independent processing unit within a CPU

★ Must-know

📌 More cores improve performance when software is designed to use them through parallel processing or multithreading.

Further detail

• Dual-core, quad-core, hexa-core, and deca-core CPUs contain 2, 4, 6, and 10 cores respectively, while CPUs can also contain 12, 16, 24, or 64 cores.

• Each CPU core can perform its own fetch-decode-execute cycle.

💡 Memory Hook

D-Q-H-D: dual, quad, hexa, deca-core.

📖 3. Cache Memory

🔑 Key Concepts & Definitions

• Cache : small, very fast memory located in or near the CPU that stores frequently accessed instructions and data

📝 Essential Points

• The cache levels are:
  • L1: smallest and fastest
  • L2: larger but slower than L1
  • L3: larger and slower than L1 and L2, faster than RAM, and shared between cores

📌 A larger cache reduces the need to fetch information from slower main memory (RAM), reducing average memory access time and CPU waiting time.

💡 Memory Hook

L1 → L2 → L3: increasing size and decreasing speed.

📖 4. Parallel Processing

🔑 Key Concepts & Definitions

• Parallel processing : the processing of multiple instructions or tasks at the same time

★ Must-know

📌 Parallel processing can significantly improve performance in software designed to use multiple cores.

Further detail

• Parallel processing is particularly useful for multitasking and workloads that can be divided into separate tasks, such as processing different parts of a video-rendering task simultaneously.

💡 Memory Hook

Different CPU cores simultaneously handling separate pieces of a video-rendering task.

📖 5. Threading Models

🔑 Key Concepts & Definitions

• Threading : how a processor handles and executes multiple sequences of instructions called threads
• Single threading : allows a CPU or core to execute one thread at a time
• Multilevel threading : allows a single CPU core to manage more than one thread at a time

📝 Essential Points

📌 Multilevel threading can improve performance in multitasking and parallel processing.

💡 Memory Hook

Single threading handles one thread at a time, whereas multilevel threading handles more than one thread at a time.

📖 6. Combined Performance and Bottlenecks

🔑 Key Concepts & Definitions

• Bottleneck : when one CPU characteristic limits the performance of the rest of the CPU

★ Must-know

• The combined CPU performance factors are:
  • clock speed
  • number of cores
  • cache size
  • parallel processing
  • threading

Further detail

• A CPU with high clock speed, many cores, and a small cache may access RAM more frequently, causing the CPU to wait for data and reducing overall performance.

💡 Memory Hook

A weak characteristic, such as a small cache, can limit otherwise strong CPU characteristics.

📊 Synthesis Tables

Cache Levels

Level	Size	Speed and sharing
L1	Smallest	Fastest
L2	Larger than L1	Slower than L1
L3	Larger than L1 and L2	Slower than L1 and L2, faster than RAM, shared between cores