A re-ranking technique, called "PageRank" brings a successful story behind the Google™ search engine. Many studies focus on finding an efficient way to compute the PageRank scores of a large web graph. Researchers propose to compute them sequentially by reducing the I/O cost of disk access, improving the convergence rate, or even employing Peer-2-Peer architecture, etc. However, only a few concentrate on computation using parallel processing techniques. In this paper, we propose a Partition-based parallel PageRank algorithm that can be efficiently run on a low-cost parallel environment like PC cluster. For comparison, we also study other two well-known PageRank techniques, and provide an analytical discussion of their performance in terms of I/O and synchronization cost, as well as memory usage. Experimental results show a promising improvement on a large artificial web graph synthesized from the TH domain.