Transport triggered architecture (TTA) has been shown to provide an efficient way to design application specific instruction set processors. However, the interconnection network of TTA is based on simple-bus, which consumes much extra power for specific data transport. In this paper, we employ segmented-bus to solve this problem. How to partition the buses lies on the placement of macro blocks, and the manual creation of an optimal placement would be a labor-intensive process, leading to high design costs. Instead, we propose automatic approach to place the macro blocks specialized to given applications. For several real life cryptographic applications, a factor of