[Help-gsl] Numerical integration of sharply peaked function

[Jorge Fernandez de Cossio Diaz]

I have a function of a single variable with a very sharp peak at a known 
position. I want to integrate this function in a finite interval.

I tried using qags, but it underflows often. I think that the adaptive sampling 
it is missing the sharp peak. How can I tell GSL where is the sharp peak, so 
that it knows before hand that it should sample in its vicinity? Also, if the 
peak is VERY sharp, I want GSL to keep sampling until it finds the very narrow 
neighborhood of the peak where the function is distinct from zero.

How can I do this?

Thanks in advance.
---------------------------------------------------

[http://5.cim.co.cu/cim.gif]