Results 1 to 6 of 6

Thread: Crummy KVM performance with virtio, what's wrong?

  1. #1
    Join Date
    Sep 2006
    Location
    Brooklyn
    Beans
    10
    Distro
    Xubuntu 8.10 Intrepid Ibex

    Crummy KVM performance with virtio, what's wrong?

    I've configured a server to support our development team and test a potential configuration for production.

    I initially tried to configure Xen only do discover Dom0 is not supported in 8.10, but KVM looked like an attractive alternative. After some trail and error I got a VM working with virtio_net and virtio_blk with the image on an LVM2 volume.

    While I'm unimpressed with the net IO and CPU results, the disk IO is less than desirable. Looking around the discussion board I came across this post

    http://ubuntuforums.org/showthread.p...ghlight=virtio

    Which leads me to believe something isn't quite right with my configuration. I've attached the XML dumps from libvirt for vm01 (file based image) and vm02 (lvm based image). I've also attached performance benchmarks from the host system and both VMs. While running the benchmark on the host I shutdown both VMs. While benchmarking the VMs I disabled the other.

    If anyone out there has any insight, it would be greatly appreciated. I would love to get this working and deploy KVM over VMWare ESX.

    colossus (Host)
    Code:
    ---------------
    1x1000BASE-T, 8 Core, 16GB RAM, 4x15K RPM Hardware RAID5
    ---------------
    Linux colossus 2.6.27-9-server #1 SMP Thu Nov 20 22:56:07 UTC 2008 x86_64 GNU/Linux
    ---------------
    vendor_id       : GenuineIntel
    cpu family      : 6
    model           : 23
    model name      : Intel(R) Xeon(R) CPU           X5460  @ 3.16GHz
    stepping        : 6
    cpu MHz         : 3158.737
    cache size      : 6144 KB
    physical id     : 0
    siblings        : 4
    core id         : 0
    cpu cores       : 4
    apicid          : 0
    initial apicid  : 0
    fpu             : yes
    fpu_exception   : yes
    cpuid level     : 10
    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 lm constant_tsc arch_perfmon pebs bts rep_good nopl pni monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr dca sse4_1 lahf_lm
    bogomips        : 6317.47
    clflush size    : 64
    cache_alignment : 64
    address sizes   : 38 bits physical, 48 bits virtual
    power management:
    *******************************************************************
    me@colossus:~$ iperf -c xxx.xxx.xxx.14 -r -t 60 -w 256K
    ------------------------------------------------------------
    Server listening on TCP port 5001
    TCP window size:   256 KByte (WARNING: requested   256 KByte)
    ------------------------------------------------------------
    ------------------------------------------------------------
    Client connecting to xxx.xxx.xxx.14, TCP port 5001
    TCP window size:   256 KByte (WARNING: requested   256 KByte)
    ------------------------------------------------------------
    [  5] local xxx.xxx.xxx.206 port 43906 connected with xxx.xxx.xxx.14 port 5001
    Waiting for server threads to complete. Interrupt again to force quit.
    [ ID] Interval       Transfer     Bandwidth
    [  5]  0.0-60.0 sec  6.55 GBytes    938 Mbits/sec
    
    me@colossus:/data/vm$ bonnie++ -s 32g -n 256
    Version 1.03c       ------Sequential Output------ --Sequential Input- --Random-
                        -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
    Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
    colossus        32G 76730  88 218667  46 96284  17 76906  81 273038  23  1141   3
                        ------Sequential Create------ --------Random Create--------
                        -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
                  files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                    256 67646  87 +++++ +++ 13106  15 72814  92 +++++ +++ 12249  16
    colossus,32G,76730,88,218667,46,96284,17,76906,81,273038,23,1141.5,3,256,67646,87,+++++,+++,13106,15,72814,92,+++++,+++,12249,16
    
    me@colossus:~/nbench-byte-2.2.3$ ./nbench 
    
    BYTEmark* Native Mode Benchmark ver. 2 (10/95)
    Index-split by Andrew D. Balsa (11/97)
    Linux/Unix* port by Uwe F. Mayer (12/96,11/97)
    
    TEST                : Iterations/sec.  : Old Index   : New Index
                        :                  : Pentium 90* : AMD K6/233*
    --------------------:------------------:-------------:------------
    NUMERIC SORT        :          1499.7  :      38.46  :      12.63
    STRING SORT         :          330.64  :     147.74  :      22.87
    BITFIELD            :      6.0249e+08  :     103.35  :      21.59
    FP EMULATION        :           327.6  :     157.20  :      36.27
    FOURIER             :           33763  :      38.40  :      21.57
    ASSIGNMENT          :          45.146  :     171.79  :      44.56
    IDEA                :          9100.4  :     139.19  :      41.33
    HUFFMAN             :          3057.6  :      84.79  :      27.07
    NEURAL NET          :          67.226  :     107.99  :      45.43
    LU DECOMPOSITION    :          2112.7  :     109.45  :      79.03
    ==========================ORIGINAL BYTEMARK RESULTS==========================
    INTEGER INDEX       : 109.365
    FLOATING-POINT INDEX: 76.847
    Baseline (MSDOS*)   : Pentium* 90, 256 KB L2-cache, Watcom* compiler 10.0
    ==============================LINUX DATA BELOW===============================
    CPU                 : 8 CPU GenuineIntel Intel(R) Xeon(R) CPU           X5460  @ 3.16GHz 3159MHz
    L2 Cache            : 6144 KB
    OS                  : Linux 2.6.27-9-server
    C compiler          : gcc version 4.3.2 (Ubuntu 4.3.2-1ubuntu11) 
    libc                : libc-2.8.90.so
    MEMORY INDEX        : 28.018
    INTEGER INDEX       : 26.758
    FLOATING-POINT INDEX: 42.622
    Baseline (LINUX)    : AMD K6/233*, 512 KB L2-cache, gcc 2.7.2.3, libc-5.4.38
    * Trademarks are property of their respective holder.
    vm01 (Guest)

    Code:
    <domain type='kvm'>
      <name>vm01</name>
      <uuid>7efa39f0-1c1c-a8c7-8af8-8ac93852e455</uuid>
      <memory>1048576</memory>
      <currentMemory>1048576</currentMemory>
      <vcpu>1</vcpu>
      <os>
        <type>hvm</type>
        <boot dev='hd'/>
      </os>
      <features>
        <acpi/>
      </features>
      <clock offset='utc'/>
      <on_poweroff>destroy</on_poweroff>
      <on_reboot>restart</on_reboot>
      <on_crash>destroy</on_crash>
      <devices>
        <emulator>/usr/bin/kvm</emulator>
        <disk type='file' device='disk'>
          <source file='/data/vm/1/vm01-kvm/disk0.qcow2'/>
          <target dev='hda' bus='ide'/>
        </disk>
        <disk type='file' device='disk'>
          <source file='/data/vm/1/vm01-kvm/disk1.qcow2'/>
          <target dev='hdb' bus='ide'/>
        </disk>
        <interface type='bridge'>
          <mac address='52:54:00:e7:40:74'/>
          <source bridge='br0'/>
          <target dev='vnet1'/>
          <model type='virtio'/>
        </interface>
        <input type='mouse' bus='ps2'/>
        <graphics type='vnc' port='-1' listen='127.0.0.1'/>
      </devices>
    </domain>
    *******************************************************************
    me@vm01:~$ iperf -c xxx.xxx.xxx.14 -r -t 60 -w 256K
    ------------------------------------------------------------
    Server listening on TCP port 5001
    TCP window size:   256 KByte (WARNING: requested   256 KByte)
    ------------------------------------------------------------
    ------------------------------------------------------------
    Client connecting to xxx.xxx.xxx.14, TCP port 5001
    TCP window size:   256 KByte (WARNING: requested   256 KByte)
    ------------------------------------------------------------
    [  5] local xxx.xxx.xxx.208 port 42280 connected with xxx.xxx.xxx.14 port 5001
    Waiting for server threads to complete. Interrupt again to force quit.
    [ ID] Interval       Transfer     Bandwidth
    [  5]  0.0-60.0 sec  1.03 GBytes    147 Mbits/sec
    
    me@vm01:~$ bonnie++ -s 2g -n 256
    Version 1.03c       ------Sequential Output------ --Sequential Input- --Random-
                        -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
    Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
    vm01             2G  3867  90 15525  69 15619  46  4908  98 54912  78 337.1  97
                        ------Sequential Create------ --------Random Create--------
                        -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
                  files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                    256  2957  95 36378  91  2908  80  2953  95 57011  96  2443  72
    vm01,2G,3867,90,15525,69,15619,46,4908,98,54912,78,337.1,97,256,2957,95,36378,91,2908,80,2953,95,57011,96,2443,72
    
    me@vm01:~/nbench-byte-2.2.3$ ./nbench 
    
    BYTEmark* Native Mode Benchmark ver. 2 (10/95)
    Index-split by Andrew D. Balsa (11/97)
    Linux/Unix* port by Uwe F. Mayer (12/96,11/97)
    
    TEST                : Iterations/sec.  : Old Index   : New Index
                        :                  : Pentium 90* : AMD K6/233*
    --------------------:------------------:-------------:------------
    NUMERIC SORT        :           333.4  :       8.55  :       2.81
    STRING SORT         :          16.104  :       7.20  :       1.11
    BITFIELD            :      5.3435e+07  :       9.17  :       1.91
    FP EMULATION        :           25.21  :      12.10  :       2.79
    FOURIER             :          2502.4  :       2.85  :       1.60
    ASSIGNMENT          :          6.9528  :      26.46  :       6.86
    IDEA                :          595.95  :       9.11  :       2.71
    HUFFMAN             :          303.15  :       8.41  :       2.68
    NEURAL NET          :          4.4887  :       7.21  :       3.03
    LU DECOMPOSITION    :          195.32  :      10.12  :       7.31
    ==========================ORIGINAL BYTEMARK RESULTS==========================
    INTEGER INDEX       : 10.474
    FLOATING-POINT INDEX: 5.922
    Baseline (MSDOS*)   : Pentium* 90, 256 KB L2-cache, Watcom* compiler 10.0
    ==============================LINUX DATA BELOW===============================
    CPU                 : AuthenticAMD QEMU Virtual CPU version 0.9.1 3159MHz
    L2 Cache            : 512 KB
    OS                  : Linux 2.6.27-9-server
    C compiler          : gcc version 4.3.2 (Ubuntu 4.3.2-1ubuntu11) 
    libc                : libc-2.8.90.so
    MEMORY INDEX        : 2.446
    INTEGER INDEX       : 2.747
    FLOATING-POINT INDEX: 3.284
    Baseline (LINUX)    : AMD K6/233*, 512 KB L2-cache, gcc 2.7.2.3, libc-5.4.38
    * Trademarks are property of their respective holder.
    
    me@vm01:~$ lsmod | grep virtio
    virtio_net             19584  0 
    virtio_pci             15104  0 
    virtio_ring            12800  1 virtio_pci
    virtio                 14084  2 virtio_net,virtio_pci
    vm02 (Guest)

    Code:
    <domain type='kvm'>
      <name>vm02</name>
      <uuid>2a4a9234-e8cf-b01c-c21c-b0aea193fc4b</uuid>
      <memory>1048576</memory>
      <currentMemory>1048576</currentMemory>
      <vcpu>1</vcpu>
      <os>
        <type>hvm</type>
        <boot dev='hd'/>
      </os>
      <features>
        <acpi/>
      </features>
      <clock offset='utc'/>
      <on_poweroff>destroy</on_poweroff>
      <on_reboot>restart</on_reboot>
      <on_crash>destroy</on_crash>
      <devices>
        <emulator>/usr/bin/kvm</emulator>
        <disk type='file' device='disk'>
          <source file='/dev/virtualization/vm02'/>
          <target dev='vda' bus='virtio'/>
        </disk>
        <interface type='bridge'>
          <mac address='52:54:00:fd:e2:84'/>
          <source bridge='br0'/>
          <target dev='vnet2'/>
          <model type='virtio'/>
        </interface>
        <input type='mouse' bus='ps2'/>
        <graphics type='vnc' port='-1' listen='127.0.0.1'/>
      </devices>
    </domain>
    *******************************************************************
    me@vm02:~$ iperf -c xxx.xxx.xxx.14 -r -t 60 -w 256K
    ------------------------------------------------------------
    Server listening on TCP port 5001
    TCP window size:   256 KByte (WARNING: requested   256 KByte)
    ------------------------------------------------------------
    ------------------------------------------------------------
    Client connecting to xxx.xxx.xxx.14, TCP port 5001
    TCP window size:   256 KByte (WARNING: requested   256 KByte)
    ------------------------------------------------------------
    [  5] local xxx.xxx.xxx.209 port 40649 connected with xxx.xxx.xxx.14 port 5001
    Waiting for server threads to complete. Interrupt again to force quit.
    [ ID] Interval       Transfer     Bandwidth
    [  5]  0.0-60.0 sec    887 MBytes    124 Mbits/sec
    
    me@vm02:~$ bonnie++ -s 2g -n 256
    Version 1.03c       ------Sequential Output------ --Sequential Input- --Random-
                        -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
    Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
    vm02             2G  4226  92 17541  71 22417  53  5650  98 80662  70  1283  97
                        ------Sequential Create------ --------Random Create--------
                        -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
                  files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                    256  3453  95 37876  98  3839  83  3598  95 53105  98  3534  77
    vm02,2G,4226,92,17541,71,22417,53,5650,98,80662,70,1282.5,97,256,3453,95,37876,98,3839,83,3598,95,53105,98,3534,77
    
    me@vm02:~/nbench-byte-2.2.3$ ./nbench
    
    BYTEmark* Native Mode Benchmark ver. 2 (10/95)
    Index-split by Andrew D. Balsa (11/97)
    Linux/Unix* port by Uwe F. Mayer (12/96,11/97)
    
    TEST                : Iterations/sec.  : Old Index   : New Index
                        :                  : Pentium 90* : AMD K6/233*
    --------------------:------------------:-------------:------------
    NUMERIC SORT        :          337.04  :       8.64  :       2.84
    STRING SORT         :          15.984  :       7.14  :       1.11
    BITFIELD            :      5.3752e+07  :       9.22  :       1.93
    FP EMULATION        :          25.219  :      12.10  :       2.79
    FOURIER             :          2389.3  :       2.72  :       1.53
    ASSIGNMENT          :           6.939  :      26.40  :       6.85
    IDEA                :          597.85  :       9.14  :       2.71
    HUFFMAN             :          302.46  :       8.39  :       2.68
    NEURAL NET          :          4.5689  :       7.34  :       3.09
    LU DECOMPOSITION    :          193.12  :      10.00  :       7.22
    ==========================ORIGINAL BYTEMARK RESULTS==========================
    INTEGER INDEX       : 10.487
    FLOATING-POINT INDEX: 5.843
    Baseline (MSDOS*)   : Pentium* 90, 256 KB L2-cache, Watcom* compiler 10.0
    ==============================LINUX DATA BELOW===============================
    CPU                 : AuthenticAMD QEMU Virtual CPU version 0.9.1 3159MHz
    L2 Cache            : 512 KB
    OS                  : Linux 2.6.27-9-server
    C compiler          : gcc version 4.3.2 (Ubuntu 4.3.2-1ubuntu11) 
    libc                : libc-2.8.90.so
    MEMORY INDEX        : 2.443
    INTEGER INDEX       : 2.755
    FLOATING-POINT INDEX: 3.241
    Baseline (LINUX)    : AMD K6/233*, 512 KB L2-cache, gcc 2.7.2.3, libc-5.4.38
    * Trademarks are property of their respective holder.
    
    me@vm02:~/nbench-byte-2.2.3$ lsmod | grep virtio
    virtio_net             19584  0 
    virtio_blk             14728  4 
    virtio_pci             15104  0 
    virtio_ring            12800  1 virtio_pci
    virtio                 14084  3 virtio_net,virtio_blk,virtio_pci
    Anything look odd to you?

    Thanks!
    EightBit
    Last edited by EightBitHustler; December 31st, 2008 at 03:23 AM.

  2. #2
    Join Date
    Sep 2006
    Location
    Brooklyn
    Beans
    10
    Distro
    Xubuntu 8.10 Intrepid Ibex

    Re: Crummy KVM performance with virtio, what's wrong?

    I came across this page:
    http://libvirt.org/drvqemu.html

    KVM hypervisor: The driver will probe /usr/bin for the presence of qemu-kvm and /dev/kvm device node. If both are found, then KVM fullyvirtualized, hardware accelerated guests will be available.
    I've installed the proper kvm & qemu packages as prescribed in the documentation, but I don't see /dev/kvm.

    Code:
    dportello@colossus:~$  ls /dev/k*
    /dev/kmem  /dev/kmsg
    Could this be an issue?

  3. #3
    Join Date
    Sep 2006
    Location
    Brooklyn
    Beans
    10
    Distro
    Xubuntu 8.10 Intrepid Ibex

    Re: Crummy KVM performance with virtio, what's wrong?

    I also came across this post:
    http://ubuntuforums.org/showthread.p...37#post2553459

    Which recommends:
    Code:
    sudo modprobe kvm-intel
    My results are:
    Code:
    FATAL: Error inserting kvm_intel (/lib/modules/2.6.27-9-server/kernel/arch/x86/kvm/kvm-intel.ko): Operation not supported
    I see that these kvm modules exist:
    Code:
    meo@colossus:~$ ls /lib/modules/2.6.27-9-server/kernel/arch/x86/kvm/
    kvm-amd.ko  kvm-intel.ko  kvm.ko
    I see these modules are loaded:
    Code:
    me@colossus:~$ lsmod | grep kvm
    kvm                   165744  0
    Could the lack of a properly loaded kvm-intel module be the cause?

  4. #4
    Join Date
    Sep 2006
    Location
    Brooklyn
    Beans
    10
    Distro
    Xubuntu 8.10 Intrepid Ibex

    Re: Crummy KVM performance with virtio, what's wrong?

    ...And I finally answered my own question by rereading the docs and diving further into the forum

    according to...
    https://wiki.ubuntu.com/kvm

    Typing dmesg you may find the following at the end:-
    kvm: disabled by bios
    Sure enough...
    Code:
    me@colossus:~$ dmesg | grep kvm
    [   27.586229] kvm: disabled by bios
    [  282.772155] kvm: disabled by bios
    DOH!

    Moral of the story RTFM!

  5. #5
    Join Date
    Sep 2006
    Location
    Brooklyn
    Beans
    10
    Distro
    Xubuntu 8.10 Intrepid Ibex

    Re: Crummy KVM performance with virtio, what's wrong?

    This morning I had someone in my office enable the virtualization extensions in the BIOS. I re-ran my bonnie benchmarks and here are the results

    Code:
    -------------------------------------------------------------------------------
    Host system
    -------------------------------------------------------------------------------
    
    Version 1.03c       ------Sequential Output------ --Sequential Input- --Random-
                        -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
    Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
    colossus        32G 76730  88 218667  46 96284  17 76906  81 273038  23  1141   3
                        ------Sequential Create------ --------Random Create--------
                        -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
                  files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                    256 67646  87 +++++ +++ 13106  15 72814  92 +++++ +++ 12249  16
    colossus,32G,76730,88,218667,46,96284,17,76906,81,273038,23,1141.5,3,256,67646,87,+++++,+++,13106,15,72814,92,+++++,+++,12249,16
    
    -------------------------------------------------------------------------------
    Before virtual extensions were enabled
    -------------------------------------------------------------------------------
    
    Version 1.03c       ------Sequential Output------ --Sequential Input- --Random-
                        -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
    Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
    vm02             2G  4226  92 17541  71 22417  53  5650  98 80662  70  1283  97
                        ------Sequential Create------ --------Random Create--------
                        -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
                  files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                    256  3453  95 37876  98  3839  83  3598  95 53105  98  3534  77
    vm02,2G,4226,92,17541,71,22417,53,5650,98,80662,70,1282.5,97,256,3453,95,37876,98,3839,83,3598,95,53105,98,3534,77
    
    -------------------------------------------------------------------------------
    After virtual extensions were enabled
    -------------------------------------------------------------------------------
    
    Version 1.03c       ------Sequential Output------ --Sequential Input- --Random-
                        -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
    Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
    vm02             2G 71855  85 260611  25 211285  31 88313  93 414699  30 +++++ +++
                        ------Sequential Create------ --------Random Create--------
                        -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
                  files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                    256 67351  90 517207  98 44379  72 71398  96 +++++ +++ 34968  72
    vm02,2G,71855,85,260611,25,211285,31,88313,93,414699,30,+++++,+++,256,67351,90,517207,98,44379,72,71398,96,+++++,+++,34968,72
    As you can see the performance is near the host system. On this test I had another vm running, but not doing anything with it.

    On some of the metrics, the performance appears to be better than the host. I'm sure this could be due to caching.

    Net performance
    Code:
    -------------------------------------------------------------------------------
    Host system
    -------------------------------------------------------------------------------
    
    me@colossus:~$ iperf -c guardian -r -t 60
    ------------------------------------------------------------
    Server listening on TCP port 5001
    TCP window size: 85.3 KByte (default)
    ------------------------------------------------------------
    ------------------------------------------------------------
    Client connecting to guardian, TCP port 5001
    TCP window size: 16.0 KByte (default)
    ------------------------------------------------------------
    [  5] local xxx.xxx.xxx.206 port 53483 connected with xxx.xxx.xxx.205 port 5001
    Waiting for server threads to complete. Interrupt again to force quit.
    [ ID] Interval       Transfer     Bandwidth
    [  5]  0.0-60.0 sec  6.58 GBytes    942 Mbits/sec
    
    -------------------------------------------------------------------------------
    Guest system
    -------------------------------------------------------------------------------
    me@vm02:~$iperf -c guardian -r -t 60        
    ------------------------------------------------------------
    Server listening on TCP port 5001
    TCP window size: 85.3 KByte (default)
    ------------------------------------------------------------
    ------------------------------------------------------------
    Client connecting to guardian, TCP port 5001
    TCP window size: 16.0 KByte (default)
    ------------------------------------------------------------
    [  5] local xxx.xxx.xxx.209 port 38786 connected with xxx.xxx.xxx.205 port 5001
    Waiting for server threads to complete. Interrupt again to force quit.
    [ ID] Interval       Transfer     Bandwidth
    [  5]  0.0-60.0 sec  5.10 GBytes    731 Mbits/sec
    CPU performance
    Code:
    -------------------------------------------------------------------------------
    Host system (8 cores, I'm guessing nbench only operates on one core)
    -------------------------------------------------------------------------------
    me@colossus:~/nbench-byte-2.2.3$ ./nbench 
    
    BYTEmark* Native Mode Benchmark ver. 2 (10/95)
    Index-split by Andrew D. Balsa (11/97)
    Linux/Unix* port by Uwe F. Mayer (12/96,11/97)
    
    TEST                : Iterations/sec.  : Old Index   : New Index
                        :                  : Pentium 90* : AMD K6/233*
    --------------------:------------------:-------------:------------
    NUMERIC SORT        :          1503.3  :      38.55  :      12.66
    STRING SORT         :          333.84  :     149.17  :      23.09
    BITFIELD            :       6.047e+08  :     103.73  :      21.67
    FP EMULATION        :           327.2  :     157.01  :      36.23
    FOURIER             :           33823  :      38.47  :      21.61
    ASSIGNMENT          :          45.084  :     171.55  :      44.50
    IDEA                :          9100.4  :     139.19  :      41.33
    HUFFMAN             :          3067.7  :      85.07  :      27.16
    NEURAL NET          :          67.506  :     108.44  :      45.62
    LU DECOMPOSITION    :            2115  :     109.57  :      79.12
    ==========================ORIGINAL BYTEMARK RESULTS==========================
    INTEGER INDEX       : 109.622
    FLOATING-POINT INDEX: 77.026
    Baseline (MSDOS*)   : Pentium* 90, 256 KB L2-cache, Watcom* compiler 10.0
    ==============================LINUX DATA BELOW===============================
    CPU                 : 8 CPU GenuineIntel Intel(R) Xeon(R) CPU           X5460  @ 3.16GHz 3159MHz
    L2 Cache            : 6144 KB
    OS                  : Linux 2.6.27-9-server
    C compiler          : gcc version 4.3.2 (Ubuntu 4.3.2-1ubuntu11) 
    libc                : libc-2.8.90.so
    MEMORY INDEX        : 28.130
    INTEGER INDEX       : 26.788
    FLOATING-POINT INDEX: 42.722
    Baseline (LINUX)    : AMD K6/233*, 512 KB L2-cache, gcc 2.7.2.3, libc-5.4.38
    * Trademarks are property of their respective holder.
    -------------------------------------------------------------------------------
    Guest system (1 core)
    -------------------------------------------------------------------------------
    me@vm02:~/nbench-byte-2.2.3$ ./nbench 
    
    BYTEmark* Native Mode Benchmark ver. 2 (10/95)
    Index-split by Andrew D. Balsa (11/97)
    Linux/Unix* port by Uwe F. Mayer (12/96,11/97)
    
    TEST                : Iterations/sec.  : Old Index   : New Index
                        :                  : Pentium 90* : AMD K6/233*
    --------------------:------------------:-------------:------------
    NUMERIC SORT        :          1497.8  :      38.41  :      12.62
    STRING SORT         :          334.08  :     149.28  :      23.11
    BITFIELD            :      6.0398e+08  :     103.60  :      21.64
    FP EMULATION        :           326.4  :     156.62  :      36.14
    FOURIER             :           33739  :      38.37  :      21.55
    ASSIGNMENT          :          45.026  :     171.33  :      44.44
    IDEA                :          9096.4  :     139.13  :      41.31
    HUFFMAN             :            3060  :      84.85  :      27.10
    NEURAL NET          :          66.933  :     107.52  :      45.23
    LU DECOMPOSITION    :          2119.4  :     109.80  :      79.28
    ==========================ORIGINAL BYTEMARK RESULTS==========================
    INTEGER INDEX       : 109.453
    FLOATING-POINT INDEX: 76.798
    Baseline (MSDOS*)   : Pentium* 90, 256 KB L2-cache, Watcom* compiler 10.0
    ==============================LINUX DATA BELOW===============================
    CPU                 : GenuineIntel QEMU Virtual CPU version 0.9.1 3159MHz
    L2 Cache            : 2048 KB
    OS                  : Linux 2.6.27-9-server
    C compiler          : gcc version 4.3.2 (Ubuntu 4.3.2-1ubuntu11) 
    libc                : libc-2.8.90.so
    MEMORY INDEX        : 28.113
    INTEGER INDEX       : 26.728
    FLOATING-POINT INDEX: 42.595
    Baseline (LINUX)    : AMD K6/233*, 512 KB L2-cache, gcc 2.7.2.3, libc-5.4.38
    * Trademarks are property of their respective holder.
    Final verdict, KVM is fantastic!
    Last edited by EightBitHustler; December 31st, 2008 at 05:28 PM. Reason: more info

  6. #6
    Join Date
    Jan 2009
    Beans
    3

    Re: Crummy KVM performance with virtio, what's wrong?

    z
    Last edited by cderigo; January 28th, 2009 at 10:18 PM.

Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •