Semi-direct algorithms for the MP2 energy and gradient

✍ Scribed by Michael J. Frisch; Martin Head-Gordon; John A. Pople

• Publisher: Elsevier Science
• Year: 1990
• Tongue: English
• Weight: 777 KB
• Volume: 166
• Category: Article
• ISSN: 0009-2614
• DOI: 10.1016/0009-2614(90)80030-h

✦ Synopsis

The cost (via the number of two-electron integral evaluations) and the maximum size of a direct second-order Mooller-Plesset (MP2) energy or gradient calculation are both determined by the available computer memory. Therefore we formulate semidirect MP2 methods that utilize disk space (which is usually much larger than memory size) for the steps that require most storage. In terms of the molecular basis set size, they require as little as quadratic memory and cubic disk. The amount of input/ output transfer between memory and disk is quartic plus the cost of transpositions, which is between quartic and quintic. A variety of calculations are presented comparing the fully direct, semi-direct and conventional algorithms. The semi-direct methods are shown to be superior to conventional algorithms despite requiring less disk space, and are also often preferred over the direct