In the investigation we try to compare by use of a parallel algorithm called kombat in protein DNA level to match DNA strand in the code area by two collimator connector OpenMP (of the share memory type) and MPI(of the distributed memory type). But main goal of this study is that which one offers less time for matching two DNA strand of code area on protein Level. Materilals and Methods: For measuring results it has been used of NM-009791strand with 9363 long neocleotide and XM-422197 strand with 9078 long neocleotide (obtained by NCBI web site). Results: Presented results in this article has been achieved by running parallel kombat in C language and by use of collimator connector MPI on a cluster contains 2 systems that every system has a Quad-core processor with computing power of (Intel. Core i7-3770T) and a memory of (DDR3 4GB). At the time of program running on a processor , run –time of algorithm 2705,2 and its speed-up is one, it is also program run- time with 8 processors 385,8 and its speed–up is 7. while presented results has been achieved by use of collimator connector OpenMP on a server with two processor sockets that have two Quad-core processors with computing power of (Intel. Core i7-3770T) with memory of ( 8GB DDR3). When program is running on a processor , the run –time of algorithm 2454,6 and its speed –up is one, and also the run-time of program with 8 processors 340,8 and its speed-up is 7,2. Conclusion: In the final conclusion, with comparing achieved results , the collimator connector OpenMP has less run-time and more speed–up than MPI .But in terms of hardware its hard to increasing the number of processors and also the responsibility of right access order to memory is on programmer directly. Finally the cost of constructing computers with shared memory and high number processors is very high , that these are not as so in the collimator connector MPI that is a distributed memory relation.

Keywords