Abstract-This paper presents an iterative geometric mean decomposition (IGMD) algorithm for multiple-input-multipleoutput (MIMO) wireless communications. In contrast to the conventional geometric mean decomposition (GMD) algorithm, the proposed IGMD does not require the explicit Kth root computation in the preprocessing stage but depends on a carefully constructed iterative procedure that generates the GMD in its limit. We prove analytically that the proposed IGMD is guaranteed to converge to the exact GMD under certain sufficient conditions, and propose three different constructions achieving this condition. Both numerical simulations and complexity analysis of the proposed IGMD have been conducted and compared with the conventional GMD. Simulation results show that our new IGMD algorithm effectively reduces the complexity overhead and hence is more advantageous for low-complexity implementations.

One technique to increase the yield of memories is to incorporate spare rows columns into the main memory array. However, due to the long bit-line and line lengths in today's SOC technology, it is not efficient to replace faulty cells with a spare row or a spare column. Therefore, the redundant rows (columns) are divided into row (column) blocks, respectively. The reconfiguration is performed at the row (column) block level instead of the conventional row (column) level. We first propose a redundancy analysis algorithm suitable for built-in implementation or embedded memories with 2-D extended local repair-most algorithm, Due to the efficient utilization of redundant the manufacturing yield and reliability can be improved significantly. Thereafter, a simulator is implemented with language for evaluating the repair rates of the proposed algorithm. Simulation results show the proposed approach can increase the repair rate and fabrication yield significantly.

No abstract

No abstract

No abstract