deltasig68

@Pic55 Did you try, yet? Any results?

Thank you TK, I did have a look at the link and followed through the write up. I got curious to see what the OPi PC had for these parameters since I run many times big projects on that on including the "make -j2" that crashed the H5 and the OPI One. /proc/sys/vm/overcommit_memory=0, /proc/sys/vm/overcommit_ratio=50. The H5 that run the 4.10 had one more parameter /proc/sys/vm/overcommit_kbytes=0 (no idea what that does) and the swapiness for all is "0". The system on the OPi PC is: 3.4.113-sun8i #10 SMP PREEMPT Thu Feb 23 19:55:00 CET 2017 armv7l armv7l armv7l GNU/Linux. It run without any problem with memory usage getting as high as 80% and the swap (default swap of Armbian) got to 85%. I'll play around with the parameters to see if something changes. I have a question off this topic but I think you are the right person to ask: are the voltage/frequency parameters adjusted for the specific board before hand or do I have to go and fine tune? the H5 runs a bit on the "hot" side without much to do! Thank you.

About a year ago I reported the kswapd0 running out of control on OPi PC and Igor, after couple of tries made it behave. Today, I got OPi One loaded with the latest 5.26 Ubuntu D/T and the OPi PC2 with today's daily Ubuntu D/T and both demonstrated the following behaviour when pressed for memory: Specifically, running a "make -j2", when the available RAM got below 38000, the kswapd0 started getting more cpu time and eventually everything stopped waiting for disk I/O (80% wa displayed in "top"). Interestingly, not a single byte of swap storage was used. I tried different combinations of programs and everything worked fine until the demands on RAM got the available RAM below 38000. The sure fire combination were a long make (with just one thread) and FF displaying the finance page of yahoo! This behaviour is, I dare say, identical to the one I encounter a year ago. Any ideas? Please, let me know if you need any more information.