Mmmh. Looks like something else is working in a wrong way, meaning: It's not due to the latest Project C commits. I've cross tested with a v5.9-r1 BMQ on a 5.9.15 kernel and a v5.9-r3 PDS on 5.9.14 and the problem persists each time. I'd re-test with an older kernel to see if the difference in behaviour is in the in-kernel development.
Anyway, my question from above remains for now: Does someone of you have an 'ideal' setting of "sched_timeslice" for an old dual-core Intel without HT? (The 6000 worked well for many months.)
Hi Alfred,
ReplyDeleteI'm not conviced of your intended improvements.
Either kernel 5.9.14 itself or your commits introduce in-browser video-playback stuttering in BMQ mode -- maybe in network throughput.
Can you please have a look at it and possibly related code changes?
Anyway, many great thanks for your work!
BR, Manuel
Or do I get an advice from you to alter my sched_timeslice=6000 setting?
DeleteBR, Manuel
Mmmh. Looks like something else is working in a wrong way, meaning: It's not due to the latest Project C commits. I've cross tested with a v5.9-r1 BMQ on a 5.9.15 kernel and a v5.9-r3 PDS on 5.9.14 and the problem persists each time. I'd re-test with an older kernel to see if the difference in behaviour is in the in-kernel development.
DeleteAnyway, my question from above remains for now:
Does someone of you have an 'ideal' setting of "sched_timeslice" for an old dual-core Intel without HT? (The 6000 worked well for many months.)
TIA and best regards,
Manuel