Add advice on location and how this affects the GPT results
This question will likely keep coming up about where to run the GPT in relation to the server it's testing and what that means for results.
Task is to come to a recommendation for users and add this to the docs. Could go a few ways on this. Do we recommend they test as close as possible at the full throughput or test further away but with expectations set that there will be a network latency delay.