The first equation that we use extensively is attached in the MathML form. It is a sum of exponents, I am not sure that GP supports it directly but who knows. The N can vary from 24 to 1024, so it puts a computational burden on our simulator.

The second one, which we do not use but might use in the future, is a dot product operation. This operation, in fact, is partly supported by Intel processors with the SSE extensions. The problem that we have 144 elements from which the dot product is calculated, so SSE is not the best solution here However, we can split 144 elements into blocks and calculate each of them with a help of SSE. It is not very complicated to do even with inline Assembly language.