I had the same problem. I found a good solution:
In the end I used the "Gaussian-Peak-Fit.vi" . It has an input "parameter bounds"; this has to be set to the 4 parameters that define your desired gauss function (center, amplitude, width, offset).
The parameter bounds needs a max and a min-value for each parameter. I just set max=min. (see diagram: 4 values in and 8 values out as a cluster).
The data feed into the Y-input is irrelevant. It is only necessary so that the gauss-fit routine does not complain.
The inspiration for this solution came from the following statement in the help of the Gaussian-Peak-Fit.vi:
"If you know the exact value of certain parameters, you can set the lower and upper bounds of the parameters equal to the known values."
Here is a nice programm:
