I have a set of ROOT macros to process
TTrees and fill histograms (
TH1D). Lately I observed that the jobs I send to my local institute’s cluster stall or work very slowly. I was told that usage of
/dev/random might affect this as the strange slow-down only occurs on the nodes where multiple instances of my code run.
I have several questions on searching a solution of my issue:
- Can compiling my code help (not using macros directly)?
- How to reduce the use of
Note: I’m not using any random numbers. I’m talking about core ROOT code that utilises
If I recall /dev/random is used in a feature of the interpreter that prevent dereference of bad pointers.
Indeed, compiling your code (even just via ACliC) should solve the problem.
Thank you for prompt reply. I will try compiling my code and report back the results.
I compiled my code (executable, not ACliC) and I still see
/dev/random in the process file descriptors. Is this expected but it will not be used?
Yes, it is still setup and ready to be used if/when the interpreter is invoked.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.