<html><head></head><body>What are you settling that fixes the Nvidia experience complaint? <br><br><div class="gmail_quote">On October 4, 2016 9:15:56 AM PDT, Hristo Iliev <hristo@hiliev.eu> wrote:<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<pre class="k9mail">Hi Martin,<br /><br />Am 04.10.2016 10:09, schrieb Martin Schrodt:<br /><blockquote class="gmail_quote" style="margin: 0pt 0pt 1ex 0.8ex; border-left: 1px solid #729fcf; padding-left: 1ex;"> Hi Hristo,<br /> <br /><blockquote class="gmail_quote" style="margin: 0pt 0pt 1ex 0.8ex; border-left: 1px solid #ad7fa8; padding-left: 1ex;"> No need to sleep/wake - my X99-based system starts with TSC disabled:<br /> <br /> $ dmesg | grep TSC<br /> [ 0.000000] tsc: Fast TSC calibration using PIT<br /> [ 0.077986] TSC deadline timer enabled<br /> [ 0.203383] TSC synchronization [CPU#0 -> CPU#1]:<br /> [ 0.203384] Measured 974558547804462 cycles TSC warp between CPUs,<br /> turning off TSC clock.<br /> [ 0.203388] tsc: Marking TSC unstable due to check_tsc_sync_source<br /> failed<br /> <br /> Consequently, tsc is not among the clock sources listed in<br /> available_clocksource. KVM is not happy about that:<br /> <br /> [16739.200656] kvm: SMP vm cre!
ated on
host with unstable TSC; guest <br /> TSC<br /> will not be reliable<br /></blockquote> <br /> Ok, so an X99-board that behaves like this even on a fresh start.<br /> Interesting.<br /> <br /><blockquote class="gmail_quote" style="margin: 0pt 0pt 1ex 0.8ex; border-left: 1px solid #ad7fa8; padding-left: 1ex;"> But I haven't observed any instabilities of the Windows 10 guest, <br /> which<br /> happily runs with 4 virtual CPUs (2 virtual hyperthreaded CPUs) bound <br /> to<br /> two cores of my i7-5820K.<br /></blockquote> <br /> This really makes me think there's something else involved in this<br /> behaviour. Maybe the CPU configuration (I use "Skylake-Client") exposes<br /> TSC to the guest, so if you put that on, it'll use it?<br /> <br /> Can you check what kind of virtual CPU you use?<br /> <br /></blockquote><br />I use 'passthrough', otherwise the default 'Haswell-noTSX' that <br />virt-manager picks results in NVIDIA GeForce Experience complaining <br />about unrecogn!
ised CPU
type, so the TSC gets pretty much exposed to the <br />guest.<br /><br />Just found this paste with a boot log from another system with the same <br />motherboard, though with an older BIOS version and a Xeon E5-1620 v3 <br />CPU:<br /><br /><a href="https://pastelink.net/fjm">https://pastelink.net/fjm</a><br /><br />It shows the same problem with unsynchronised TSCs. Could also be <br />something specific to the LGA 2011v3 processors or to the EFI BIOS. I'll <br />take a look in the BIOS options when time permits.<br /><br /><blockquote class="gmail_quote" style="margin: 0pt 0pt 1ex 0.8ex; border-left: 1px solid #729fcf; padding-left: 1ex;"> Cheers,<br /> Martin<br /></blockquote><br />Cheers,<br />Hristo<br /><br /><hr /><br />vfio-users mailing list<br />vfio-users@redhat.com<br /><a href="https://www.redhat.com/mailman/listinfo/vfio-users">https://www.redhat.com/mailman/listinfo/vfio-users</a><br /></pre></blockquote></div></body></html>