2.4 Basic Measures of Computer Performance
COS2621 - Computer Organisation · 2. Performance Issues
Performance Issues
Basic Measures of Computer Performance
Performance measures are essential for evaluating how well a computer system operates. These measures help in comparing different systems and understanding their capabilities. The most common performance measures include throughput, latency, and response time.
Clock Cycles
One increment, or pulse, of a clock is referred to as a clock cycle. The clock cycle is crucial in determining how quickly a computer can process instructions. Each instruction typically takes several clock cycles to execute, depending on its complexity. For example, if a processor operates at a clock speed of 2 GHz, it can perform 2 billion clock cycles per second.
Throughput
Throughput refers to the number of tasks a computer can complete in a given time period. It is usually measured in tasks per second. For example, if a computer can process 300 tasks in 5 minutes, its throughput is calculated as follows:
Throughput = 300 tasks / 5 minutes = 60 tasks/minuteHigher throughput indicates better performance. This metric is particularly important in environments where many tasks need to be processed simultaneously.
Latency
Latency is the time taken to complete a single task. It is often measured in milliseconds (ms). For instance, if a task takes 150 ms to complete, the latency is 150 ms. Latency is crucial in applications where immediate feedback is necessary, such as online gaming or video conferencing.
Response Time
Response time is the total time taken from when a request is made until the first response is received. This includes processing time and any delays caused by input/output operations. For example, if a user clicks a button and the system responds after 0.8 seconds, the response time is 0.8 seconds.
Calculating Performance Metrics
To better understand these concepts, consider a scenario where a computer processes 600 tasks in 15 minutes. First, calculate the throughput:
Throughput = 600 tasks / 15 minutes = 40 tasks/minuteIf the average latency for these tasks is 100 ms, this means each task takes approximately 0.1 seconds to complete. Remember to distinguish between throughput and latency, as confusing the two can lead to incorrect conclusions about system performance.
Benchmarking
Benchmarks are standard tests used to measure the performance of computer systems. They provide a way to compare the performance of different systems under similar conditions. Common benchmarks include SPEC (Standard Performance Evaluation Corporation) benchmarks, which measure CPU performance, and TPC (Transaction Processing Performance Council) benchmarks, which evaluate database performance.
Example of Benchmarking
If a system scores 1600 on a SPEC benchmark, this means it can process tasks at a rate that is 60% faster than a system scoring 1000. Understanding these benchmarks helps in making informed decisions when purchasing or upgrading computer systems.
Summary
- Throughput measures how many tasks a computer can complete in a given time.
- Latency measures the time taken to complete a single task.
- Response time measures the total time from request to first response.
- Benchmarks are used for comparing system performance.
Check your understanding
- What is a clock cycle and why is it important?
- How would you calculate the throughput of a system if it completed 720 tasks in 12 minutes?
- What does a high response time indicate about a computer system?
- Why are benchmarks important in evaluating computer performance?
Test yourself
4 quick questions on this lesson. Pick an answer to see if you're right.
1. If a system scores 1500 on a SPEC benchmark, what does this indicate?
2. What does throughput measure in a computer system?
3. What is the primary purpose of benchmarking in computer performance evaluation?
4. If a computer has a latency of 200 ms, how long does it take to complete a single task?