Hello All, Our application requires that the guest clock should be synch’ed to the host clock with the accuracy of 10’s of nano-sec. of course we can achieve this accuracy when working with physical hosts and running PTP. The question is what accuracy can be achieved with virtual machines. In my tests I only try to sync the guest to the host clock using the kvm-clock – and I don’t run any protocol for network clock sync (NTP/PTP) My future plan is to run PTP on physical hosts and to sync the guests to their host’s clock. 1. My VM is Fedora 21 and the host is Fedora 22 (kernels are 4.1 and 4.2 accordingly) 2. My Intel core processor supports the constant-tsc. 3. The default clock in the guest is kvm-clock . When I run a test to measure packet delay between the host and the VM (which normally should be about 5 to 10 micro-sec) - I see about 1 second delay I see that delta time between the VM and host is about 1 sec - which means that the sync. Accuracy is bad What accuracy do I have to expect with this clocking mode (kvm-clock) ? tens of ms/us/ns? Bigger ? Maybe I’ve missed some configuration to achieve a better accuracy ? when we get a good accuracy – I’m going to run PTP on the host – and expect that the VM clock will be in-sync. Here’s the 1st segment of /proc/cpuinfo: processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 63 model name : Intel(R) Xeon(R) CPU E5-2690 v3 @ 2.60GHz stepping : 2 microcode : 0x31 cpu MHz : 1200.062 cache size : 30720 KB physical id : 0 siblings : 24 core id : 0 cpu cores : 12 apicid : 0 initial apicid : 0 fpu : yes fpu_exception : yes cpuid level : 15 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm ida arat epb pln pts dtherm tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm xsaveopt cqm_llc cqm_occup_llc bugs : bogomips : 5199.72 clflush size : 64 cache_alignment : 64 address sizes : 46 bits physical, 48 bits virtual power management: Thanks and Regards, Avi -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
