From My experience, To perform SVAR analyis, both Eviews and Stata are fine. However, You may not able to produce error band using Eviews, which is pretty easy to do it using Stata. But I don't like the way how stata deal with time series data.
P. S. It is very difficult to publish your paper or answer external examiner if you do not produce error band for your IRF. It is like you run regression without reporting standard error or t-statistics.
I prefere RATS to others for time series analysis.