A review of cacheSweave (0.4-4)
Sweave for HPC
cacheSweave is the most elegant way of incorporating large computations in Sweave documents; the results of cached chunks are stored in a stashR database, making it possible to make large computations once on a cluster and then edit the document recompiling it many times on a thin client. In a same time, it does not require any changes in code chunks, so it is completely transparent. I haven’t encountered any problems while using this package; what may be an issue is that the package does not fetch the side effects of functions called inside cached chunks – for instance the console output or graphics; however this may be omitted in most cases by making additional, non-cached chunks.
