This dissertation focuses on improving the division-by-convergence algorithm. While the division by convergence algorithm has many advantages, it has some drawbacks, such as a need for extra bits in the multiplier and a large ROM table for the initial approximation. To mitigate these problems, two new methods are proposed here. In addition, the research scope is extended to seek an efficient architecture for implementing a divider with Quantum-dot Cellular Automata (QCA), an emerging technology. For the first proposed approach, a new rounding method to reduce the required precision of the multiplier for division by convergence is presented. It allows twice the error tolerance of conventional methods and inclusive error bounds. The proposed method further reduces the required precision of the multiplier by considering the asymmetric error bounds of Goldschmidt dividers. The second proposed approach is a method to increase the speed of convergence for Goldschmidt division using simple logic circuits. The proposed method achieves nearly cubic convergence. It reduces the logic complexity and delay by using an approximate squarer with a simple logic implementation and a redundant binary Booth recoder. Finally, a new architecture for division-by-convergence in QCA is proposed. State machines for QCA often have synchronization problems due to the long wire delays. To resolve this problem, a data tag method is proposed. It also increases the throughput significantly since multiple division computations can be performed in a time skewed manner using one iterative divider. / text