Okay, some back and forth is there in any case, but without

knowing the time distance between the individual instances it's

hard to tell whether what I'm thinking of might help.

 

> besides, I checked my former simple tester this morning after it ran for a 

> whole night, it lagged much.

 

Could you btw quantify the lagging?

// I didn't pay attention to it, but about half an hour's lagging certainly exist. 

when you will go home, you can also have a try to see result tomorrow morning.

 

Your tester sets only a single, constant rate repeatedly, right? //yes, the simplest test just setting a single 10000 down repeatly.

If so, then the thought of adjustment likely won't help.