Re: [time-nuts] Linux TSC clocksource on multi-core systems

2014-05-01 Thread Wojciech Owczarek
Laszlo, What sometimes helps is additional kernel command line parameters, namely acpi=off (maybe you wouldn't have to disable the PM settings in the BIOS if you had this) and noapic, also there is the clocksource=tsc parameter which should make TSC the preferred clock source (that's if it's

Re: [time-nuts] Linux TSC clocksource on multi-core systems

2014-04-30 Thread Jason Rabel
Laszlo, Have you tried choosing between your different available clock sources and letting NTP run for a few days and log the stats? Check: /sys/devices/system/clocksource/clocksource0/ available_clocksource current_clocksource ___ time-nuts

Re: [time-nuts] Linux TSC clocksource on multi-core systems

2014-04-29 Thread Laszlo Hanyecz
The problem isn't so much NTP, it's that the kernel can't use the TSC as the clock source, so it falls back to something like the HPET or ACPI timer which are orders of magnitude less resolution than the TSC. Here is the Atom D510 that shows the synchronization problem. Booting with 'nosmp'

Re: [time-nuts] Linux TSC clocksource on multi-core systems

2014-04-26 Thread Michael Tharp
On 04/26/2014 02:27 PM, Laszlo Hanyecz wrote: It's fine to disable the additional cores/cpus on a dedicated NTP machine, but I wonder if there is a solution that allows both the TSC and all the cores to be used at the same time. Is it even possible to completely sync the counters across CPUs