Sweave for "big" data analysis

As others have said, there are packages that provide caching.

I haven't used them, because I like to keep my projects as 
self-contained as possible:  adding a dependency on one of those 
packages is undesirable[1].  What I do in the case where there are time 
consuming calculations is to do all the calculations in a script, and 
save the results (using save()).  Then the Sweave document will load the 
objects (using load()) and do post-processing, plotting, etc.