R
's default configuration options disable memory profiling, shared-library libR
, static-library libR.a
, GNU
gettext
, valgrind
headers, and link-time optimisation.
Some of these (like valgrind
) must just be because not everyone will use it. Others might be answered elsewhere on SO (like Is there any reason why not to use link time optimization? might answer --enable-lto[=no]
).
But what is the possible disadvantage or cost of enabling memory profiling in R
?
Oops, apparently the answer can be found from https://stat.ethz.ch/R-manual/R-devel/library/utils/html/Rprofmem.html:
The memory profiler slows down
R
even when not in use, and so is a compile-time option.