VPLIB

The MVP-9500 system contained all of the scalar system components with the addition of an MVP-9500/8 printed circuit board and the VPLIB library of F80 callable subroutines it is shown schematically in Figure 7. 

The design criteria for the VPLIB library were as follows ... 

This would certainly have resulted in an operational library in the shortest possible time but at considerable sacrifice in efficiency. All of the VPLIB subroutines were therefore written entirely in Z80A/MVP-9500 assembly language and this produced modules which contained, on average, one third of the assembler instructions produced by F80 for the same operation coded in FORTRAN. In addition to these straightforward savings a considerable amount of hand optimization was possible on the assembler level subroutines. 

Fortran Code for Matrix Multiplication using VPLIB... 

In most intstances however merely being able to arrange the calculation as a series of vector operations, without worrying over the "unit address increment" requirement, makes extremely good if not maximal use of AFPP, VP or AP. As an illustration of this point Table XIII shows "normal" FORTRAN code for a pivotal condensation matrix inverter (the author is unfortunately by now anonymous) and Table XIV shows the vectorized version for the MVP-9500 at about two thirds completion. The VPLIB version is completely (as far as the author can manage at leastl) vectorized and written in assembler. Most of the vectorization is fairly obvious and only the reduction loops contain any obscurity. In order to maintain peak vector efficiency the MVP-9500 reduction loop does a little more work than is strictly necessary an alternative would ruin the vector flow. It is left as an exercise to the determined reader to unravel the full correspondence between Tables XIII and XIV. 

Fortran/VPLIB Code for Matrix Matrix "Outer Product"... 

Fortran/VPLIB Code for Gauss-Jordan Matrix Inverter... 

It was mentioned earlier that a number of special purpose routines, which do not appear in the VPLIB index, have been developed for use in structural chemistry. The most frequent requirements encountered in this area are those concerned with molecular geometry and, more specifically, with the calculation of interatomic distances, angles and torsion angles. These geometric quantities are best evaluated by vector algebra and this will always involve the calculation of vector components, lengths, direction cosines, vector cross products and vector dot products. Attention should therefore be directed at the best possible way of implementing the calculations described in the latter list on the MVP-9500. 

Interatomic distances, angles and torsion angles may be efficiently calculated with the routines discussed above plus the remainder of VPLIB. A number of other special purpose routines have been developed and are under development but these will not be discussed here. 

The calculation of bond lengths etc. by means of VPLIB calls has already been discussed. In the case of the energies, the same four algebraic expressions are used over and over again with... 

The construction of an MVP-9500 version of the molecular mechanics program which ran faster than the author s PDP-11/40 (FIS) version was an undemanding if tedious process but, as with construction of VPLIB routines, arriving at an efficient rather than brute force coding will take several iterations. 

---