TDD can’t help you when it comes to determining the performance of a software development project.
Imagine if you will a situation where you are dealing with a 3rd party vended application, and you work through all of the functional requirements in a UAT environment. Imagine that you have previous experience where related software projects have performed by orders of magnitude better in PROD than in UAT. When the important software development project is migrated to Production, how will it actually perform?
Anyone with any decent amount of experience already knows the horrific result of finding out that it performs orders of magnitude worse.
There is no easy answer when it comes to this, but it’s a fact of life. What is important is having metrics for each key component of the project so that you can at least pinpoint where in Production things are worse than what they should be. And then do a lot of guessing.
From an end user’s standpoint, perfectly functioning functionality that performs significantly slower than expected can be as bad, if not worse, than broken functionality.
Keep this in mind when developing any new software.