It is common in many applications to leverage all kind of objects that are used as data. The types of data can be primitive, object reference, and collection types. F# itself has its own unique types, especially collection types.
These types, especially when doing many computing-intensive operations, are crucial. Optimizations need not only relate to the location of the value of data stored according to the types, but also how we access them can have a big impact on overall performance.
The deciding factors determining the best types to use, besides the memory locations, can vary. This is why it is quite hard to measure qualitatively. Measuring quantitatively can be tricky, although the statistical numbers may be informative.
We shall use the knowledge gained in Chapter 2, Performance Measurement relating to the qualitative understanding of the internals of running F# code (the IL, tooling, and the GC) and have quantitative measures (the execution durations...