GRU training time compared to LSTM

Question

I am using GRU and LSTM algorithms to analyse some network traffic dataset, I often get GRU training time longer than LSTM.

Is this normally ?