rootbench
                                
                                
                                
                                    rootbench copied to clipboard
                            
                            
                            
                        Implement an option to globally set/fix the number of parallel processing units
I would like to have a way to globally set/fix the number of processing units the parallel benchmaks will run on.
Some expert reviews of this potential feature:
"We already have the RB_TMP_DIR env variable, we could add another to control the argument passed to EnableImplicitMT" - @eguiraud
"we also have the rootbench datadir as a variable! (...) these vars can be easily exposed through cmake in the ADD_BENCHMARK_FOO functions" - @stwunsch
Can you elaborate on the use case?