Sorry for being late on this weekly report (again).
I merge the weekly report 11 & 12 in this one :
After 2 week on this problem of complete node splitting i finally found that the problem was not in my code (i even rewrite it completely due to some problem )
but the problem was in the under step made by OpenDht (and by the DHT).
If you want to see the little line that bug my code for too much week.
I also correct the layout of the graph output by the benchmark since i thought my code wasn't working (and it was on TODO )
I also speed up the initialization of search :
My PR is also up and i will push another PR this week for the multidimensional key search on the PHT.
Anyway, see you next week