Parallel Computing for Data Science

50 CHAPTER 3. SCHEDULING

The function do1pset() itself is fairly straightforward. Note that one of

the components of the object returned by the call to the regression function

lm() and then summary() is adj.r.squared, the adjusted R

2

value.

The end result will be that the call to dochunk() with psetsstart equal

to 1 will return rows 1 and 2 of the final output seen in Section 3.4.3.

Thus chunking is handled in this manner, in spite of the lack of a chunking

capability in snow itself.

That’s quite a bit to digest! The partitioning of work due to chunking was

rather intricate, and a nonchunked version would have been much simpler.

But we will find in Section ...