SOS: Why IRFs are not smooth? So many kinks there

I do not know what’s wrong. Anybody can give me some hints?
I have searched across the forum, somebody said there could be model problems. But I have used checked it few times, they are correct.

I have used the model_diagnostics command but nothing appears wrong.
attached is the mod file and initial values for endogenous variables
thanks a lot!

IRFs look like this:
http://www3.nd.edu/~xli21/untitled.jpg

I have solved the problem by one of my professors. He solved a big problem for me!
The problem here is that I have too big standard_deviations for the shocks.

Use the pruning option of stoch_simul and shock sizes won’t matter.