Quantum information processing and its associated technologies has reached an interesting and timely stage in their development where many different experiments have been performed establishing the basic building blocks. The challenge moving forward is to scale up to larger sized quantum machines capable of performing tasks not possible today. This raises a number of interesting questions like: How big will these machines need to be? how many resources will they consume? This needs to be urgently addressed. Here we estimate the resources required to execute Shors factoring algorithm on a distributed atom-optics quantum computer architecture. We determine the runtime and requisite size of the quantum computer as a function of the problem size and physical error rate. Our results suggest that once experimental accuracy reaches levels below the fault-tolerant threshold, further optimisation of computational performance and resources is largely an issue of how the algorithm and circuits are implemented, rather than the physical quantum hardware