Change compilation settings to add in symbol settingsA default R installation typically has certain compiler settings placed in the equivalent of the C:\Program Files\R\R-3.3.1\etc\x64\Makeconf that strips information needed for profiling during the Rcpp compilation process, in particular a line which reads: DLLFLAGS=-s . To override this and add some additionally needed flags, one should add a folder and file to their home directory which overrides and appends necessesary compilation flags. To a file located at a location equivalent to C:\Users\YOURNAME\.R\Makevars on your machine (note the ‘.’ before R), add the following lines:
-gdwarf-2appears in the compilation messages, and that
-sis missing in the final linker step.
Run a profiler which understands MinGW compiled codeThe next key step is to run a profiler which can understand the Unix like symbols on windows. Two free and good options are Very Sleepy and AMD’s code analyst (which also works on Intel chips). Very Sleepy is very good at basic timings and providing stack traces, while AMD’s profiler is able to drill down to the assembly of a process. Both profilers are good but an example with AMD is shown below.
- Open the program and setup a quick session to start and run a sample R script that uses your code, such as in the example shown below.
- Next run the profiler and get ready to look at results. For example, here I can see that half the time was spent in my code, versus half in the R core’s code (generating random numbers)And digging further down I can see at the assembly level what the biggest bottlenecks were in my code
Thanks for sharing this. Do you happen to have an update for this? AMD’s code analyst does not exist anymore and I fail to get it running with Very Sleepy (it runs the code but does not resolve the symbols so the output is useless).