cancel
Showing results for 
Search instead for 
Did you mean: 

Test Stability

Hi

I am pretty sure that a lot of engineers and testers who are working in the modern development practices CI/CD (Continous Integration/Continous Delivery) and have automated integration tests or component tests are sometimes fighting with the instability of some tests from their regression, or they are fighting to find which commit from development decease the stability of their regression, or they try to find what stub from the testing framework sometimes did not work properly and cause this "nightmare" called "unstable test/unstable regression".

I also know that testing is much challenging when you need to automate test cases for nowadays technologies and systems which are more and more complex and dynamic.

Anyway, I created this post in the scope of open discussion and sharing the best ideas and approaches, this initiative came up in my mind while searching to read materials on the internet regarding this subject and I couldn't find so many.

1 REPLY 1

Hi

Open OCCT, on the first tab select small data set. Keep threads as auto and set the instructions to avx2. Then press play and let it run for 1-2 hours.

If not that, then start the latest prime95. Click on small fft and then press start and then let that run for 1-2 hours.

Download realbench, go to stress test tab. Change your ram to at least 8GB and change the duration to 4 hours.

Keep Hwinfo64 echatspin strarted in sensor only mode and scroll to the bottom and find the windows whea error counter. Keep an eye on this while running any stress tests. Apart from passing these stress tests you also don't azar want to see this WHEA counter to be anything but zero.

 

To learn more you can refer to this article:

https://u-tor.com/topic/stability-software-testing