The timer T3-RESPONSE holds the maximum wait time for a response of a request message.
T3-timer has not as such defined values, but there is a method to count the T3-response timer value.
Actually it means that after the request message GTP-U peer will start the timer and if peer is not getting the response within that timer expires, it has to send request message again.
The T3-RESPONSE timer shall be started when a signalling request message (for which a response has been defined) is sent. A signalling message request or response has probably been lost if a response has not been received before the T3-RESPONSE timer expires. At the expiry of the timer the request is re-transmitted if the total number of request attempts is less than N3-REQUESTS times.
Now if i come to your query how can we calculate that timer value.
We know a GTP-U peer may send an Echo Request on a path to the other GTP-U peer to find out if it is alive. Echo Request messages may be sent for each path in use. So from here it will find out the shortest path of other GTP-U peer and will note down the time that in how much time it has got the Echo response . This time will be the T3-timer value. So for the next request message that will path will be the dedicated path and T3-timers will expect the response with in that time