Hello,
I am not sure if this is the right place to start this topic.
I am running the example mH-recoil analysis from FCCAnalyses. Starting with release 2021-05-26 I start to get trouble when processing large amounts of data (i.e running ZZ + WW + ZH):
About to run process p8_ee_ZH_ecm240 with 10000000 events
done
*** Error in `python': double free or corruption (fasttop): 0x000000000b6248e0 ***
followed by a very long backtrace. Running in 2021-05-23 does not show this problem at all.
hello @nmorange ,
I guess you have removed the build and install directories and recompiled after sourcing a different environment? we changed root version around that date, so could be an issue if you’re not aligned.
Clement
and I’m moving this as FCCAnalysis topic
@nmorange, is your issue fixed?
Hello @clement.helsens . The issue seems to have disappeared. I am quite certain that I had recompiled the code, but maybe I did not do it properly.
Another thing to notice is that running takes quite a bit of RAM, so maybe there was an issue on this at some point.
Anyway, thanks a lot for your answers, and I can close the topic.
Very good @nmorange!
About the memory, see this thread I opened a while back:
We never really went to the bottom of it, as anyway I decide to use the batch for large processing.
Please feel free to jump in if you find out something interesting!