Shulou(Shulou.com)06/01 Report--

In view of how PerfDog helps to automate performance testing, this article introduces the corresponding analysis and solutions in detail, hoping to help more partners who want to solve this problem to find a simpler and easier way.

Necessity of performance testing

The number of performance issues in the phases of the entire project

Performance problems didn't happen at the beginning, nor did they pop up one day, but accumulated with our development progress.

And the same problem, the same person, the experience and time it takes to deal with it at different times are completely different.

So the performance problem seems to be the technical problem of the R & D team, but in essence it is the development process problem of the R & D team.

If we can standardize the process and make sure that each version has a data display, once a problem is found and dealt with in time, then the optimization time can be greatly reduced; and the performance of each version of manpower is more chicken ribs, so it can be handled in an automated way, so whether the automatic operation will have an impact on the performance data we get, let's explore below.

The influence of Automation on Application performance data comparison of the first set of tests

Test background:

1. Open Perfdog and record the performance data of manual running and automatic running.

two。 The automation function used this time is Airtest.

Test case:

1. Airtest IDE connection is not enabled, manual running function

two。 Open Airtest IDE connection and run manually

3. Open Airtest IDE connection and use automatic script running function

4. Disconnect Airtest IDE connection

5. Close the Airtest IDE process

Automation script:

Will only run a small combat function for a short period of time

The disconnection of the following test case refers to:

Let's take a look at FPS first.

Obviously, we found that whether or not to run games in an automated way has almost no effect on FPS.

Let's take a look at memory.

It is found that automation has no effect on memory, and whether automation is turned on or not is almost the same for memory.

Let's take a look at CPU.

We found that Total cpu usage increased significantly when opening IDE connections to airtest, and Total cpu usage increased when running automated scripts. On the other hand, the cpu usage of app has little effect.

This is because when opening the connection to airtest ide, ide uses minicap service to get screenshots of the phone, so it will affect the overall utilization of cpu, while airtest needs image search and matching when running scripts, so it also takes up cpu. However, there will be no impact on app usage.

Comparison of the second group of tests

This test is not suitable for automated scripts. Compare the impact of ide alone.

Test case:

1. Static pages do not connect to airtest ide

two。 Static page connection airtest ide

3. The static page is disconnected from airtest ide and does not exit ide.

4. Still page disconnects airtest ide connection and exits ide

FPS data

Whether IDE is enabled or not has no effect on the fps of the application.

Memory

Memory doesn't matter.

CPU utilization rate

Like the conclusion of the first group, enabling ide will have an impact on total cpu utilization. It should be noted that performance consumption still exists after disconnecting the IDE from the phone, because the mincap service is actually not interrupted, and the IDE cpu will not return to normal until you exit and disable it.

Third set of data

The choice is the mobile phone APP, not the game.

FPS

Memory

CPU

We found that the conclusion is the same as above.

It is recommended to use normalized CPU utilization

Why is this recommended as a measure of CPU usage? because it is found that normalization is more suitable for automation and more accurate documentation about normalized utilization:

Introduction to standardized utilization

It is entirely possible to obtain the performance data of an application in an automated way, because the data we obtained are all for a single application, so the automated operation will not be calculated within the application. However, if you connect to the automated sdk, you should consider separately. The resources consumed by SDK will be counted on the application head.

This is the answer to the question about how PerfDog helps automate performance testing. I hope the above content can be of some help to you. If you still have a lot of doubts to be solved, you can follow the industry information channel to learn more about it.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.