var::Data performance test shouldn't use rand() (stress test should use rand)

tyler-gilbert commented 6 years ago

@shomagan Did you see this issue?

I haven't looked closely at how the stress and performance tests are done but the idea is:

Performance should run a fixed number of cases with fixed data sizes so that when the test is run on different platforms they can be compared
Stress should run random cases with random sizes many, many times so that we can see if the system will fail

I have found a bunch of bugs in both the API and the OS that I have been working on thanks to the tests you have created so far. Thanks!

shomagan commented 6 years ago

Yes, I read the READMY file, in the previous tests it is possible to use rand() for the performance test, I will correct where necessary.How do I understand the most voluminous test should be stress test?

tyler-gilbert commented 6 years ago

Yes. The stress test should be the biggest. The performance test shouldn't last more than a few seconds. That we many performance tests can be run and compared in a short period of time.

On Mon, Jun 25, 2018 at 6:00 AM shomagan notifications@github.com wrote:

Yes, I read the READMY file, in the previous tests it is possible to use rand() for the performance test, I will correct where necessary.How do I understand the most voluminous test should be stress test?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/StratifyLabs/sostest/issues/52#issuecomment-399926899, or mute the thread https://github.com/notifications/unsubscribe-auth/AF2jzWbMpFBo3Hho8Qfk3z9yKvuPX9_0ks5uANDKgaJpZM4UQuiM .

StratifyLabs / sostest

var::Data performance test shouldn't use rand() (stress test should use rand) #52