Howdy, Stranger!

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

Sign In Register

Quick Links


Categories

In this Discussion

› › Poor UnixBench results
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Poor UnixBench results

FredQcFredQc Member
in General

This is my UnixBench results from one of the best provider from LET:

   BYTE UNIX Benchmarks (Version 5.1.3)

   System: localhost: GNU/Linux
   OS: GNU/Linux -- 2.6.32-431.29.2.el6.x86_64 -- #1 SMP Tue Sep 9 21:36:05 UTC 2014
   Machine: x86_64 (x86_64)
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
   CPU 0: Intel(R) Xeon(R) CPU E5-2620 0 @ 2.00GHz (4000.1 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
   CPU 1: Intel(R) Xeon(R) CPU E5-2620 0 @ 2.00GHz (4000.1 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
   CPU 2: Intel(R) Xeon(R) CPU E5-2620 0 @ 2.00GHz (4000.1 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
   CPU 3: Intel(R) Xeon(R) CPU E5-2620 0 @ 2.00GHz (4000.1 bogomips)
          Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET
   12:16:24 up 2 days, 28 min,  1 user,  load average: 0.06, 0.02, 0.00; runlevel 3

------------------------------------------------------------------------
Benchmark Run: Fri Oct 17 2014 12:16:24 - 12:44:18
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables       18927875.8 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     2311.2 MWIPS (10.0 s, 7 samples)
Execl Throughput                                934.0 lps   (29.8 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        167943.2 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           40676.4 KBps  (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        539053.6 KBps  (30.0 s, 2 samples)
Pipe Throughput                              279807.9 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                  41104.7 lps   (10.0 s, 7 samples)
Process Creation                               1813.3 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   1384.9 lpm   (60.0 s, 2 samples)
Shell Scripts (8 concurrent)                    196.2 lpm   (60.1 s, 2 samples)
System Call Overhead                         256279.4 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   18927875.8   1621.9
Double-Precision Whetstone                       55.0       2311.2    420.2
Execl Throughput                                 43.0        934.0    217.2
File Copy 1024 bufsize 2000 maxblocks          3960.0     167943.2    424.1
File Copy 256 bufsize 500 maxblocks            1655.0      40676.4    245.8
File Copy 4096 bufsize 8000 maxblocks          5800.0     539053.6    929.4
Pipe Throughput                               12440.0     279807.9    224.9
Pipe-based Context Switching                   4000.0      41104.7    102.8
Process Creation                                126.0       1813.3    143.9
Shell Scripts (1 concurrent)                     42.4       1384.9    326.6
Shell Scripts (8 concurrent)                      6.0        196.2    327.0
System Call Overhead                          15000.0     256279.4    170.9
                                                                   ========
System Benchmarks Index Score                                         312.6

------------------------------------------------------------------------
Benchmark Run: Fri Oct 17 2014 12:44:18 - 13:12:40
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables       21635533.3 lps   (10.0 s, 7 samples)
Double-Precision Whetstone                     2366.0 MWIPS (9.7 s, 7 samples)
Execl Throughput                                902.5 lps   (29.7 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks        100628.5 KBps  (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks           27786.0 KBps  (30.1 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks        300432.2 KBps  (30.1 s, 2 samples)
Pipe Throughput                              283293.1 lps   (10.0 s, 7 samples)
Pipe-based Context Switching                  50425.8 lps   (10.1 s, 7 samples)
Process Creation                               1854.8 lps   (30.0 s, 2 samples)
Shell Scripts (1 concurrent)                   1491.6 lpm   (60.2 s, 2 samples)
Shell Scripts (8 concurrent)                    200.6 lpm   (60.6 s, 2 samples)
System Call Overhead                         254029.5 lps   (10.0 s, 7 samples)

System Benchmarks Index Values               BASELINE       RESULT    INDEX
Dhrystone 2 using register variables         116700.0   21635533.3   1853.9
Double-Precision Whetstone                       55.0       2366.0    430.2
Execl Throughput                                 43.0        902.5    209.9
File Copy 1024 bufsize 2000 maxblocks          3960.0     100628.5    254.1
File Copy 256 bufsize 500 maxblocks            1655.0      27786.0    167.9
File Copy 4096 bufsize 8000 maxblocks          5800.0     300432.2    518.0
Pipe Throughput                               12440.0     283293.1    227.7
Pipe-based Context Switching                   4000.0      50425.8    126.1
Process Creation                                126.0       1854.8    147.2
Shell Scripts (1 concurrent)                     42.4       1491.6    351.8
Shell Scripts (8 concurrent)                      6.0        200.6    334.4
System Call Overhead                          15000.0     254029.5    169.4
                                                                   ========
System Benchmarks Index Score                                         286.9

Why is it so slow ? The fact is I have 4 core access of Intel's 2620 cpu and the node is supposed to be "not oversold" ...

I have sent tickets regarding this issue and they do not find anything abnormal.

Is it time to move on ? :/

Fred

Comments

  • linuxthefishlinuxthefish Member

    What provider?

  • fileMEDIAfileMEDIA Member

    Is the cpu capped to a limit like 10% of a core or something?

  • LeeLee Veteran

    said: one of the best provider from LET

    Really depends who it is, I mean you could class G*H as one of the best, I don't know.

  • Nick_ANick_A Member, Top Host, Host Rep

    Who uses E5-2620 beyond DO? I don't think they give 4 cores for a LE VPS?

  • GoodHostingGoodHosting Member

    @Nick_A said:
    Who uses E5-2620 beyond DO? I don't think they give 4 cores for a LE VPS?

    E5-2620 ? Just about every ColoCrossing provider, and a large section of the Budget market pre-facebook dump load. (before the *5639 and *5520 servers became immediately popular over night due to the price alone.)

  • FredQcFredQc Member

    dr.server in Chicago.

  • ElChileElChile Member

    @FredQc said:
    dr.server in Chicago.

    Is it a storage VPS ?

  • FredQcFredQc Member

    @ElChile said:
    Is it a storage VPS ?

    Ressource pool.

    Thanked by 1ElChile
  • GreenHostBoxGreenHostBox Member
    edited October 2014

    @FredQc said:

    Seems to be geared more towards a storage VPS so you can't really expect much performance out of it. The node shouldn't be oversold because it is a XEN VPS.

  • ElChileElChile Member

    Thanks @FredQc I wondered how that resource pool would work out ...

  • drserverdrserver Member, Host Rep

    @FredQc I gave you answer in the ticket.

    If you allow i will re post it here.

  • LeeLee Veteran

    @drserver said:
    If you allow i will re post it here.

    That sounded loaded.

  • drserverdrserver Member, Host Rep

    @W1V_Lee Nothing fancy really, just private conversation which leads to following conclusion. It is 2ghz core, you are getting thread, that thread is not only yours. It is impossible to get full thread performance on virtualized environment. And as bottom line i have offered full refund if @FredQc is unhappy with service level or service quality.

    That is about it.

    Now if we are speaking if this is good result, platform is designed as backup system so cpu is not representative but it is usable for most backup scenarios.

    Thanked by 1Lee
  • tr1ckytr1cky Member

    Is this XEN? XEN usually gives poor unixbench results but has better real-world performance. I see this happening on the new iwstack XEN VMs, they have about the same real-world performance as similar-priced Vultr servers regarding CPU but their unixbench scores suck.

  • FredQcFredQc Member

    @drserver said:
    FredQc I gave you answer in the ticket.

    If you allow i will re post it here.

    No problem.

    Hi Andrej,

    I did the UnixBench test on a OpenVz platform with the exact same cpu and cores allowance and got 2413.6 pts ! (BuyVM)

    So this is how Xen works... I see it is not tuned for performance.

    Anyway, thanks for answering my questions with such details, this is really appreciated. You offer great service and support but Xen is not for me :) Refund is not necessary at this point, thanks for the offer. I will cancel the Ressource pool plan and keep my Yearly backup plan.

    Regards,

    Fred

    --

    My last two tests shows an average of 300 pts when I should get at least 2000 pts with 4 cores of 2620's.

    Please remember that e5-2620 is 2ghz cpu and that you dont use full core but threads. In total you have 2 full cores / 4 threads at your disposal.
    Now if we do the math this way: http://www.cpubenchmark.net/cpu.php?cpu=Intel+Xeon+E5-2620+@+2.00GHz&id=1214&cpuCount=2

    Node is equiped with dual e5-2620, which gives us 24 threads in total.
    Total score with 24 threads is 12610, if we divide 12610 with 24 that is 525 per thread if only you are using it at the moment.

    I can assure you 100% that on test which will last 15-20 minutes that multiple times you will share core with someone else. System is designed on that way that single abuser (or even multiple small abusers) cannot kill cpu on the node. XEN it self is doing weight balancing so every costumer when needed is getting as much CPU space as it is available and allowable by your service plan.

    If you need more CPU power i am recommending you our AbusiveCores. If you are in any way unhappy with level or quality of service that you are receiving, you have right for full refund.

    If you need any more answers, details or explanations i am on your disposal. You can contact me directly on my skype so we can solve this situation even faster.

    Thank you

    Andrej Kolakovic
    -- dr.Server Support Europe --

    Things are now resolved.

  • FredQcFredQc Member

    @tr1cky said:
    Is this XEN? XEN usually gives poor unixbench results but has better real-world performance. I see this happening on the new iwstack XEN VMs, they have about the same real-world performance as similar-priced Vultr servers regarding CPU but their unixbench scores suck.

    With the same cpu specs and allowance ?

  • LeeLee Veteran

    @fredQc

    It's a storage VPS, forget the pointless benchmark, what does the VPS not do that is causing you concerns?

  • FredQcFredQc Member

    @W1V_Lee said:
    fredQc

    It's a storage VPS, forget the pointless benchmark, what does the VPS not do that is causing you concerns?

    It was never advertised as such...

    lowendtalk.com/discussion/33903/drserver-net-low-end-xen-resource-pool-for-7-usd-only

  • LeeLee Veteran

    I will take it then that there was in fact nothing wrong with it? it's simply that you let a benchmark dictate whether it's good or not.

    I can't see what was not advertised that you are not happy with.

    Thanked by 1Pwner
  • PwnerPwner Member

    @FredQc said:
    lowendtalk.com/discussion/33903/drserver-net-low-end-xen-resource-pool-for-7-usd-only

    It's a storage server, it's not supposed to be used for high end performance. It's supposed to be used to backup data.

Sign In or Register to comment.