| 摘要 |
<p>PROBLEM TO BE SOLVED: To cancel the increase in costs for calculation accompanying size growing of calculation objective molecules in parallel processing. SOLUTION: In a calculation method using a parallel computer having a plurality of processors, when transforming 2 electronic integration from the atomic orbital base (rs|tu) to a molecular orbital base (ab|cd), the indexes (r) and (s) of the atomic orbital to be calculated are designated to the plurality of processors to obtain a parallel calculation method for performing processing concerning all the combinations gf designated indexes R and S and the indexes (t) and (u) of the atomic orbital to be calculated in each of the plurality of processors. Thus, grain sizes are equalized to connect a large number of commodity processors under high performance, thereby costs for a high speed arithmetic computer are reduced. In addition, since an inexpensive local memory can be utilized, a large main storage area can be secured as the whole computer.</p> |
