Computing requirements, chemists

There is a trade-off between the accuracy of the calculation and the amount of computation required. In general, the more severe the approximations, the more limited is the range of applicability of the particular calculation. An organic chemist who wishes to make use of the results of MO calclulations must therefore make a judgment about the suitability of the various methods to the particular problem. The rapid increases that have occurred in computer speed and power have made the application of sophisticated methods practical for increasingly larger molecules. 

FIGURE 7 Scientific research today often requires sophisticated equipment and computers. This chemist is using an Auger electron spectrometer to probe the surface of a crystal. The data collected will allow the chemist to determine which elements are present in the surface. 

Massively parallel (multiple instruction, multiple data) computers with tens or hundreds of processors are not readily accessible to the majority of quantum chemists at the present time. However the cost of currently available hypercube machines with tens of processors (each with about the power of a VAX) is comparable to that of superminis but with up to a hundred times the power. For applications of the type discussed above the performance of a machine with as few as 32 or 64 processors would be comparable to (or perhaps even exceed) that of a single processor supercomputer. Although computer requirements currently limit QMC applications (even with effective potentials) the proliferation of inexpensive massively parallel machines could conceivably make the application of relativistic effective potentials with C C quite competitive with more conventional electronic structure techniques. 

The most popular measure of the practical performance of an algorithm is the amount of CPU time which a specified computer requires to complete a specified task, commonly a single iteration of a specified SCF procedure. While this is certainly a very appealing measure (it directly reflects an expense which is uppermost in the minds of most computational chemists), it will be extremely difficult for another worker to reproduce the timing unless the specifications are very complete. 

Ab initio calculations require a great deal of computer time and memory, because every term in the calculations is evaluated explicitly. Semiempirical calculations have more modest computer requirements, allowing the calculations to be completed in a shorter time and making it possible to treat larger molecules. Chemists generally use semiempirical methods whenever possible, but it is useful to understand both methods when solving a problem. 

Structure-generating procedures of practical value should meet three requirements. First, the procedure should be exhaustive, There are numerous examples of compounds with incorrect structural assignments in the primary literature because not all structural possibilities were considered. Computers, unlike chemists at times, have no preconceived ideas about plausible structure types and, if properly programmed, excel in accuracy in the performance of repetitive, tedious tasks. Thus, a structure generator should provide the assurance that in a given output no structure equally compatible with the input has been overlooked. Second, the program should execute efficiently, i.e., on a time scale that enhances, not detracts from, user productivity. Third, the output should be a set of nonredundant structures. 

However, better use of spectral information for more rapid elucidation of the structure of a reaction product, or of a natural product that has just been isolated, requires the use of computer-assisted structure elucidation (CASE) systems. The CASE systems that exist now are far away from being routinely used by the bench chemist. More work has to go into their development. 

Quantum mechanics gives a mathematical description of the behavior of electrons that has never been found to be wrong. However, the quantum mechanical equations have never been solved exactly for any chemical system other than the hydrogen atom. Thus, the entire held of computational chemistry is built around approximate solutions. Some of these solutions are very crude and others are expected to be more accurate than any experiment that has yet been conducted. There are several implications of this situation. First, computational chemists require a knowledge of each approximation being used and how accurate the results are expected to be. Second, obtaining very accurate results requires extremely powerful computers. Third, if the equations can be solved analytically, much of the work now done on supercomputers could be performed faster and more accurately on a PC. 

People, however, have different requirements than computers. For people— which is to say chemists in their spoken and written communications—it s best that a chemical name be pronounceable and that it be as easy as possible to assign and interpret. Furthermore, it s convenient if names follow historical precedents, even if that means a particularly well-known compound might have more than one name. People can readily understand that bromomethane and methyl bromide both refer to CftyBr. 

It was not until 1968, when Don Boyd joined ns as the second theoretical chemist in our group, that the computers at Lilly started to reach a level of size, speed, and sophistication to be able to handle some of the compntational requirements of our various evaluation and design efforts. Don bronght with him Hoffmann s EHT program from Harvard and Cornell. Dne to the length of our calculations and due to the other demands on the compnter, the best we could obtain was a one-day turnaround. 

All musical composers work with the same set of notes, but the geniuses put the notes together in an extraordinarily beautiful way. Synthetic chemists all have available to them the same elements. The successful medicinal chemist will combine atoms such that amazing therapeutic effect is achieved with the resulting molecule. The computational chemist s goal should be to help the medicinal chemist by providing information about structural and electronic requirements to enhance activity, namely, information about which regions of compound space are most propitious for exploration. 

Contemporary s Tithetic chemists know detailed information about molecular structures and use sophisticated computer programs to simulate a s Tithesis before trying it in the laboratory. Nevertheless, designing a chemical synthesis requires creativity and a thorough understanding of molecular structure and reactivity. No matter how complex, every chemical synthesis is built on the principles and concepts of general chemistry. One such principle is that quantitative relationships connect the amounts of materials consumed and the amounts of products formed in a chemical reaction. We can use these relationships to calculate the amounts of materials needed to make a desired amount of product and to analyze the efficiency of a chemical synthesis. The quantitative description of chemical reactions is the focus of Chapter 4. 

---