Hello rooters!
In our experiment, our root-files (filled with histograms) are created by a MIDAS-based analyzer. A typical root-file is found in the attachment.
If I let run the attached script, which reads in the same histo from the opened file a number of times, a memory leak of 70 MB per iteration (sums up to 3.5 GB! for iter=50) is produced.
Can anybody please tell me what is wrong here?
Taking the same code and reading in the histo “hpx” from hsimple.root does not produce a significant memory leak.
Letting the code run on “myhsimple.root”, which contains only one histo inside a TFolder, results in a minor memory leak of around 7kB per iteration.
Your help is greatly appreciated!!!
Tobias Nebel
(my apologies for the large attachment…)
myhsimple.root (379 KB)
mem_leak_1.cxx (2.24 KB)
a10999_c.root (1.94 MB)