Computer Systems Performance Modeling And Evaluation - Part 1

What is computer systems performance evaluation?

Computer performance evaluation is based upon quantifying the service delivered by a computer system. We evaluate whether the computer resources and outputs are performing at the optimal level.
For example, we might be interested in:-

• comparing the power consumption of several server farm configurations;
• knowing the response time experienced by a customer performing a reservation over the Internet;
• comparing compilers for a multiprocessor machine.

Often it is necessary to analyze only a subsystem within a larger system.

For instance, studying the performance of a certain computer system's network interface independent of the size of its memory or the type of processor.

Common goals of performance evaluation

The goals depend on the specific situation and interests and abilities of the analyst.

1. Compare alternatives:

When purchasing a new computer system,

• a pretty large number of options are available
• that impact both cost and performance, such as

            ❖ The number/types of CPUs and cores per CPU,

            ❖ Type of memory and its architecture,

            ❖ The size and number of disk drives,

            ❖ Configuration of OS and database systems etc.

• The goal of performance analysis, in this case, is to provide quantitative information about which configurations are best under specific conditions.

2. Impact analysis:

In designing new systems, or in upgrading existing systems, we need to determine the impact of adding or removing a specific feature of the system.

For instance, the designer of a new processor may have to choose between

            ❖ Adding an additional floating-point execution unit to the microarchitecture, or

            ❖ Increasing the size of the on-chip cache.

This type of analysis is often referred to as a before-and-after comparison.

3. System tuning:

To find the set of parameter values that produces the best overall performance.

• In time-shared operating systems, it is possible to control the number of processes that are allowed to actively share the processor.
• This value combined with the time quantum allocated to each process has a great impact on the perceived performance of a system.

4. Identify relative performance:

To quantify the performance of a computer system relative to

• previous generations
• competitor's systems
• customer's expectations.

Performance parameters

To measure computer performance, you have to focus on response time, latency, bandwidth, throughput, channel capacity, efficiency, availability, scalability, power consumption, compression ratio, and speed up.

1. Response time

Response time is the total amount of time taken to respond to a request. In real life, it can be compared to the time between sending a message to receiving the message.

2. Latency

Latency is basically a delay such that how much time it takes for a data packet to travel from one designated point to another. Low latency points to a positive user experience whereas high latency relates to poor user experience. In real life, you can relate with PUBG when your game lags, it's basically latency which is caused due to slow internet connection thus disrupting your gameplay.

3. Bandwidth

Bandwidth is defined as the maximum rate at which data can be transferred over an internet connection in a specific time. For example, a megabit ethernet connection has a bandwidth of 1,000 Kbps.

4. Throughput

The computer's throughput is defined as the practical measure of the amount of information successfully transmitted through a channel.

5. Speed

The term speed is usually in reference to the clock speed of the processor. It is usually measured in megahertz (MHz) or gigahertz (GHz). Computer speed is one of the leading parameters in assessing a computer's performance. For example, The Apple M1 chip has a clock speed of 3.2 GHz.

6. Channel Capacity

The channel capacity is defined to be the maximum rate at which information can be reliably transmitted through a channel.

7. Efficiency

The term efficiency refers to the peak level of performance that uses fewer inputs to achieve the highest reliable output.

8. Availability

Availability refers to less downtime which means the probability that a system will operate satisfactorily at any time.

9. Scalability

It refers to the ability of a system to be enlarged to handle large amounts of data.

10. Power consumption

The term power consumption is the electric power consumed by the system. For example, Supercomputers intake a large amount of power to handle complex instructions.

11. Compression ratio

It is the measurement of the relative reduction in resource usage by compressing data size.

You might like:

• Computer Systems Performance Modeling And Evaluation - Part 2
• Computer Systems Performance Modeling And Evaluation - Part 3
• Computer Systems Performance Modeling And Evaluation - Part 4

Thank you for reading!

This is just PART 1 of the series, I have explained the purpose of computer performance evaluation and what parameters you should focus on to improve your system performance.

If you found this article useful, feel free to go for PART 2 of this awesome blog series.

Cheers!