> On 9 Dec 2020, at 15:16, jb <[email protected]> wrote: > > > > Am 09.12.20 um 14:42 schrieb Michal Skrivanek: >> >> >>> On 9 Dec 2020, at 09:09, jb <[email protected] >>> <mailto:[email protected]>> wrote: >>> >>> Thanks Michal for the explanation. I will dive in this article and see if I >>> understand it :-). >>> >>> I use here the latest 4.4.3.12 Version. So I can hope, that newer version >>> will support my CPUs better :). >>> >>> >> >> not sure. >> 4.5 cluster level adds IceLake which is Gen10, that should work with CentOS >> 8.3 newly. >> It could be that your particular model is something in between… >> >> Either way, if you really do need every feature from the CPU and you’re >> willing to sacrifice migration compatibility you can always use CPU host >> passthrough. Then it ignores any compatibility and just uses everything it >> can(everything that KVM supports) from the underlying CPU > No I don't need every feature. My concern was mostly, if I will have any > performance downsides in normal scenarios. > A compromise makes sense then. Because on one hand you do not need compatibility with CPUs that you don’t have because they’re 10 years old, vs you’re not recompiling every application with the highest optimizations to use the latest greatest instructions…so…using whatever it detects is usually the best choice:) > >>> Am 08.12.20 um 19:12 schrieb Michal Skrivanek: >>>> qemu CPUs are mostly mapping to microarchitectures, for this one there’s >>>> excessive number of details at [1] :) >>>> >>>> but yours seems to be CofeeLake (8th gen) which is not really supported >>>> yet in that version. Well, you didn’t say anything about what your version >>>> is, so I don’t know for sure… >>>> >>>> It’s usually a compromise between what the hardware has and what has been >>>> implemented just yet >>>> >>>> Thanks, >>>> michal >>>> >>>> >>>> >>>> [1] https://en.wikichip.org/wiki/intel/microarchitectures/skylake_(server) >>>> <https://en.wikichip.org/wiki/intel/microarchitectures/skylake_(server)> >>>> >>>>> On 8 Dec 2020, at 17:09, jb <[email protected] >>>>> <mailto:[email protected]>> wrote: >>>>> >>>>> I get this output: >>>>> >>>>> "cpuFlags": >>>>> "sse4_2,hle,mpx,pti,pge,pbe,rtm,popcnt,cpuid,md-clear,lm,invtsc,invpcid_single,ibrs,tsc_deadline_timer,movbe,avx2,ibpb,pse36,umip,hypervisor,erms,fpu,bts,monitor,cmov,arch-capabilities,nx,mca,abm,pschange-mc-no,aes,ht,xsaves,ds_cpl,nonstop_tsc,adx,epb,bmi2,hwp,hwp_act_window,dtherm,aperfmperf,vme,invpcid,art,nopl,fsgsbase,pts,sep,cx8,msr,acpi,x2apic,xgetbv1,fma,flush_l1d,vmx,sse2,pat,constant_tsc,ssbd,sdbg,rdrand,clflushopt,cx16,ept,tsc_adjust,intel_pt,pse,de,stibp,sse,vpid,hwp_epp,ida,xsavec,arat,pae,clflush,tm,rdtscp,lahf_lm,cpuid_fault,pclmulqdq,fxsr,flexpriority,mtrr,syscall,ssse3,pdcm,3dnowprefetch,sse4_1,smep,rep_good,est,tpr_shadow,smap,dts,skip-l1dfl-vmentry,tm2,vnmi,hwp_notify,tsc_known_freq,mmx,dtes64,xsave,arch_perfmon,avx,rdseed,smx,ss,xtpr,f16c,bmi1,pni,pdpe1gb,apic,mce,xtopology,xsaveopt,pebs,pcid,tsc,md_clear,amd-ssbd,pln,spec_ctrl,model_Conroe,model_kvm32,model_Penryn,model_Skylake-Client-noTSX-IBRS,model_IvyBridge-IBRS,model_Broadwell-noTSX-IBRS,model_Opteron_G2,model_n270,model_SandyBridge-IBRS,model_pentium,model_kvm64,model_Westmere,model_Haswell-noTSX-IBRS,model_Haswell,model_pentium2,model_pentium3,model_Westmere-IBRS,model_Opteron_G1,model_Skylake-Client-IBRS,model_Nehalem,model_coreduo,model_Skylake-Client,model_qemu64,model_Haswell-IBRS,model_Haswell-noTSX,model_Broadwell-IBRS,model_IvyBridge,model_core2duo,model_486,model_Nehalem-IBRS,model_Broadwell-noTSX,model_SandyBridge,model_Broadwell,model_qemu32", >>>>> "cpuModel": "Intel(R) Xeon(R) E-2246G CPU @ 3.60GHz", >>>>> "cpuSockets": "1", >>>>> "cpuSpeed": "4499.377", >>>>> "cpuThreads": "12", >>>>> "deferred_preallocation": true, >>>>> >>>>> >>>>> >>>>> Does this says something to you? >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> Am 08.12.20 um 17:02 schrieb Vinícius Ferrão: >>>>>> AFAIK Client is for the i3/i5/i7/i9 families and the other one is for >>>>>> Xeon platforms. >>>>>> >>>>>> But you have pretty unusually Xeon, so it may be missing some flags that >>>>>> will properly classify the CPU. >>>>>> >>>>>> You can run this on the host to check what’s detected: >>>>>> >>>>>> [root]# vdsm-client Host getCapabilities >>>>>> >>>>>> Sent from my iPhone >>>>>> >>>>>>> On 8 Dec 2020, at 10:52, jb <[email protected]> >>>>>>> <mailto:[email protected]> wrote: >>>>>>> >>>>> _______________________________________________ >>>>> Users mailing list -- [email protected] <mailto:[email protected]> >>>>> To unsubscribe send an email to [email protected] >>>>> <mailto:[email protected]> >>>>> Privacy Statement: https://www.ovirt.org/privacy-policy.html >>>>> <https://www.ovirt.org/privacy-policy.html> >>>>> oVirt Code of Conduct: >>>>> https://www.ovirt.org/community/about/community-guidelines/ >>>>> <https://www.ovirt.org/community/about/community-guidelines/> >>>>> List Archives: >>>>> https://lists.ovirt.org/archives/list/[email protected]/message/XXI2TXLN6BW4PISDQXQGNZL5WB32CGKV/ >>>>> >>>>> <https://lists.ovirt.org/archives/list/[email protected]/message/XXI2TXLN6BW4PISDQXQGNZL5WB32CGKV/> >>>> >> > _______________________________________________ > Users mailing list -- [email protected] > To unsubscribe send an email to [email protected] > Privacy Statement: https://www.ovirt.org/privacy-policy.html > oVirt Code of Conduct: > https://www.ovirt.org/community/about/community-guidelines/ > List Archives: > https://lists.ovirt.org/archives/list/[email protected]/message/EBSLFMU5MPSZ67XHPXSZTEPKSVRMN7VN/
_______________________________________________ Users mailing list -- [email protected] To unsubscribe send an email to [email protected] Privacy Statement: https://www.ovirt.org/privacy-policy.html oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/ List Archives: https://lists.ovirt.org/archives/list/[email protected]/message/NBQ4JWI46KF5MYKIFRFBXSHRXQFTTO4L/
