Howdy, Stranger!

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

Subscribe to our newsletter

Advertise on LowEndTalk.com

Latest LowEndBox Offers

    meaning of passmark CPU benchmark
    New on LowEndTalk? Please read our 'Community Rules' by clicking on it in the right menu!

    meaning of passmark CPU benchmark

    Benchmark values from Passmark Software (https://www.cpubenchmark.net) are used by some providers to give information about cpu's.

    Is there a meaningful way to translate them into relative cpu power? For example

    Atom 230 from digicube.fr    308   (single Thread: 241)
    Atom C2750 from online.net  3800   (single Thread: 579)
    

    Is there a ~12 times difference in CPU power? If not, is there any formula to compare CPU's by using benchmark numbers?

    Comments

    • cochoncochon Member
      edited February 2017

      @yekta said:
      Is there a ~12 times difference in CPU power? If not, is there any formula to compare CPU's by using benchmark numbers?

      The 230 has just one core, the C2750 has 8 cores, also the avaton cores are much more advanced clock for clock than the older atom models. If anything I would have expected a higher difference for the multi-threaded bench, the single threaded seems reasonable for the step in architecture. Not all Atoms are equal.

      EDIT: not forgetting the huge difference in speed, 1.6 vs 2.4 GHz as well

      Thanked by 1yekta
    • No, there isn't.

      You can take those numbers as rough indicators in the ballpark. There are also far more relevant tests for many application fields (like e.g. LAMP stacks) but even those are only ballpark numbers.

      There just are way too many factors plus those are depending again on way too many other factors that can be quite different from one installation to another one.

      Thanked by 1yekta

      My favourite prime number is 42. - \forall cpu in {intel, amd, arm}: cpu->speed -= cpu->speed/100 x irandom(15, 30) | state := hacked

    • jvnadrjvnadr Member
      edited February 2017

      The benchmark number in passmark's site, is an indicator of how powerful is a cpu. This number is metered combining benchmarks has run on such cpus. It is a general indicator and is not extremely accurate always, but can show if a cpu fits for your needs. And the test had been run using specific software for comparing.
      Of course, other factors are critical for the total performance of a server, like the amount of ram, the newtork speed, the speed of the hard drive etc. But you know looking the benchmark of the cpu on what this proccessor can give to the user.
      And yes, a cpu with 3000 benchmark is multiple times quicker (in fact, can handle ~x10 times more calculations per second) than a server with 300 benchmark. What can handle, is depending of course even from the type of the software that will be used in the server.

      Thanked by 1yekta

      I am here occasionally nowadays, because I really prefer https://talk.lowendspirit.com . You should try it, it is fat-free, delicious with fresh ingredients combined with the deep knowledge of the old chefs!

    • I've found the passmark numbers to be very predictive of my own application speeds (data crunching) though there are some types of workloads they're not so great at.

      #lexit spread the word.

    • yektayekta Member
      edited February 2017

      @willie said:
      I've found the passmark numbers to be very predictive of my own application speeds (data crunching) though there are some types of workloads they're not so great at.

      they are not so great at which tasks? Web serving?

    • I dunno really. Maybe graphics? They work for my stuff so I rely on them when buying and that's worked out ok. I do wonder about multi-cpu machines, e.g. where they say a dual e5 benches at only around 1.4x a single e5. With the right kind of workload you do get almost double.

      Thanked by 1yekta

      #lexit spread the word.

    • IO speed will matter a lot as well when hosting web content. Try a few times with a big, heavy database on a hdd vs on an ssd.

      I like my uptime down low and my servers all hacked. Can see me droppin' twenty-fours with a router in the rack.
      Ya like ya Switch-Ports hot and ya servers all hacked. If ya pings real high and ya networks pitch black.

    • True. Passmark is a cpu benchmark and not about i/o. For databases, ssd is great and large ram caches also are important.

      #lexit spread the word.

    • willie said: I've found the passmark numbers to be very predictive of my own application speeds (data crunching) though there are some types of workloads they're not so great at.

      willie said: True. Passmark is a cpu benchmark and not about i/o. For databases, ssd is great and large ram caches also are important.

      Even benchmarking cpu can be very different, depending on the kind of software and the need of supporting processors like, e.g., a gpu or kuda devices for video rendering. But passmark is a good indicator on what is the power of that cpu, compared with others using the same tool.

      I am here occasionally nowadays, because I really prefer https://talk.lowendspirit.com . You should try it, it is fat-free, delicious with fresh ingredients combined with the deep knowledge of the old chefs!

    • pbgbenpbgben Member, Provider

      Its based on mathematical calculations, transcoding is not a direct match but its close enough.

      If you're seeing high cpu load then a cpu with a higher score is better. But io wait can also cause high usage.

    Sign In or Register to comment.