Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


(Netcup) RS vs. VPS
New on LowEndTalk? Please Register and read our Community Rules.

(Netcup) RS vs. VPS

neikneik Member

Hi all,

I have had a RS 1000G8 for about one year and wanted to substitute it with a VPS 1000G8 Plus.

While transferring I just noticed that the available CPU Flags differ quite a lot.
RS:

flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon rep_good nopl xtopology cpuid pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch cpuid_fault invpcid_single pti ssbd ibrs ibpb fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm mpx avx512f avx512dq rdseed adx smap clflushopt clwb avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves arat umip pku ospke

VPS:

flags : fpu de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pse36 clflush mmx fxsr sse sse2 ht syscall nx lm rep_good nopl xtopology cpuid tsc_known_freq pni pclmulqdq ssse3 cx16 pcid sse4_1 sse4_2 x2apic popcnt aes xsave hypervisor lahf_lm invpcid_single pti ssbd ibrs ibpb invpcid

As you can see no AVX flags available at all.
Does anyone know if this is the normal behaviour of the VPS lineup?

Reason I'm asking: I do some transcoding on the server so AVX would be pretty nice to have as it does have quite an impact (afaik).

Comments

  • williewillie Member

    RS supposedly has dedicated cores, VPS are shared, so if you plan to transcode a lot, you want RS.

  • neik said: Does anyone know if this is the normal behaviour of the VPS lineup?

    Yes - Netcup exposes lesser flags on the VPS lineup and there are no AVX flags unfortunately.

    For your use case, the Root Server line up is better - not only are you allowed to use the CPU constantly, but the performance will also be comparatively stable/consistent.

    If you only need high bursts for some time (when you're transcoding), it may be better to go with a separate on-demand-hourly instance that gives you better performance (there are multiple providers who have this sort of option).

    Thanked by 1uptime
  • jsgjsg Member, Resident Benchmarker

    I'm not even sure that their VPS line is KVM. It might well be containers.

  • twiigltwiigl Member
    edited June 2019

    neik said: Does anyone know if this is the normal behaviour of the VPS lineup?

    I asked netcup time ago, this is intentional, since the CPU is not dedicated, they'll be able to move to different nodes if needed. So you won't suffer any CPU features you had before. This might happen if they move your host between different CPU generations.

    jsg said: I'm not even sure that their VPS line is KVM. It might well be containers.

    Those are KVM, this CPU type is shown on VPS

    model name : QEMU Virtual CPU version 2.5+

    You can also mount .iso files there

    Thanked by 1jsg
  • Daniel15Daniel15 Member
    edited June 2019

    @twiigl said:
    Those are KVM, this CPU type is shown on VPS

    model name : QEMU Virtual CPU version 2.5+

    Note that that doesn't necessarily prove that it's KVM, for example LXC or OpenVZ within KVM could still show that, and some providers do indeed do that. And of course, someone running QEMU would see that (although why would one do that given IIRC it's full emulation?).

    If you're using systemd, you can run systemd-detect-virt to detect the virtualisation technology being used. You can detect if it's a container within a VM by running systemd-detect-virt -v to detect the VM then systemd-detect-virt -c to detect the container - if both return a value, it's a container within a VM.

    Also, most good hosts will enable host passthrough when you request it, which will show the real CPU details, including all the real flags.

    Thanked by 3saibal eva2000 uptime
  • FHRFHR Member, Host Rep
    edited June 2019

    No, they are both full KVMs.

    RS has cpu host-passthrough, VPS does not.

    Thanked by 1jsg
  • Daniel15 said: Also, most good hosts will enable host passthrough when you request it

    FHR said: RS has cpu host-passthrough

    Netcup does NOT provide (full) CPU host pass through - they only provide a limited set of flags for the RS series (and they DO NOT expose the VMX flag required for nested virtualization - they charge a pretty hefty amount to expose the VMX flag).

    Netcup is a very good/reliable/stable host though and pretty much the main complaint I have is to do with the rather limited set of CPU flags.

  • I have a server from their vserver line up, 8x vcpu, 16gb ram and 1.5tb storage.

    I've not had any issues and use it for Plex, a number of transcodes going at times too along with other tasks. Performance is great :)

  • @Daniel15 said:

    @twiigl said:
    Those are KVM, this CPU type is shown on VPS

    model name : QEMU Virtual CPU version 2.5+

    [...]

    If you're using systemd, you can run systemd-detect-virt to detect the virtualisation technology being used. You can detect if it's a container within a VM by running systemd-detect-virt -v to detect the VM then systemd-detect-virt -c to detect the container - if both return a value, it's a container within a VM.

    [...]

    This is what I get on my VPS at netcup:

Sign In or Register to comment.