Results 1 to 3 of 3
  1. #1
    Join Date
    Mar 2005
    Location
    Michigan, USA
    Posts
    523

    Centos 3.5 64bit vs Centos 4.1 64bit defferent results!

    Hi all,
    I am testing Centos 3.4 64bit and Centos 4.1 64bit on the same server and I am getting very different results. The server is a AMD athlon 64 +2800 with 1 gig of ram and a 80gig SATA 150 hard drive. If you look at the Pipe Throughput they are very different and I havent made any changes in the bios. Do you know what could be wrong? Here are the bench marks.

    ----------------------------------------------Centos 3.4----------------------------------------------
    BYTE UNIX Benchmarks (Version 4.1-wht)
    System -- Linux localhost.localdomain 2.4.21-32.0.1.EL #1 Wed May 25 14:05:46 EDT 2005 x86_64 x86_64 x86_64 GNU/Linux
    /dev/sda7 2016016 274556 1639048 15% /

    Start Benchmark Run: Thu Jun 30 00:40:30 EDT 2005
    00:40:30 up 1 min, 2 users, load average: 0.42, 0.15, 0.05

    End Benchmark Run: Thu Jun 30 00:51:47 EDT 2005
    00:51:47 up 12 min, 2 users, load average: 15.58, 6.81, 3.11


    INDEX VALUES
    TEST BASELINE RESULT INDEX

    Dhrystone 2 using register variables 376783.7 6462458.7 171.5
    Double-Precision Whetstone 83.1 1416.6 170.5
    Execl Throughput 188.3 3424.4 181.9
    File Copy 1024 bufsize 2000 maxblocks 2672.0 58308.0 218.2
    File Copy 256 bufsize 500 maxblocks 1077.0 17977.0 166.9
    File Read 4096 bufsize 8000 maxblocks 15382.0 486516.0 316.3
    Pipe Throughput 111814.6 1767585.5 158.1
    Process Creation 569.3 9829.4 172.7
    Shell Scripts (8 concurrent) 44.8 426.0 95.1
    System Call Overhead 114433.5 1713380.4 149.7
    =========
    FINAL SCORE 172.8


    --------------------------------------------------Centos 4.1------------------------------------------

    BYTE UNIX Benchmarks (Version 4.1-wht)
    System -- Linux blissa.howtowebserver.com 2.6.9-11.EL #1 Wed Jun 8 16:40:06 CDT 2005 x86_64 x86_64 x86_64 GNU/Linux
    /dev/sda7 2016016 192024 1721580 11% /

    Start Benchmark Run: Sat Jul 2 10:56:42 EDT 2005
    10:56:42 up 4 min, 2 users, load average: 0.15, 0.21, 0.09

    End Benchmark Run: Sat Jul 2 11:08:00 EDT 2005
    11:08:00 up 15 min, 2 users, load average: 15.86, 7.03, 3.22


    INDEX VALUES
    TEST BASELINE RESULT INDEX

    Dhrystone 2 using register variables 376783.7 6706845.5 178.0
    Double-Precision Whetstone 83.1 1531.8 184.3
    Execl Throughput 188.3 3679.5 195.4
    File Copy 1024 bufsize 2000 maxblocks 2672.0 66001.0 247.0
    File Copy 256 bufsize 500 maxblocks 1077.0 19557.0 181.6
    File Read 4096 bufsize 8000 maxblocks 15382.0 390164.0 253.6
    Pipe Throughput 111814.6 799836.9 71.5
    Process Creation 569.3 9983.5 175.4
    Shell Scripts (8 concurrent) 44.8 425.8 95.0
    System Call Overhead 114433.5 1662184.8 145.3
    =========
    FINAL SCORE 162.2
    -----------------------------------------------------------------------------------------------------------

  2. #2
    Join Date
    Dec 2002
    Location
    Egypt
    Posts
    151
    I have the same problem with fedora box running kernel 2.6.10-1.771_FC2smp
    I've upgraded the kernel to 2.6.12.5 and run the unixbench again without any changes
    and got a low value for Pipe Throughput

    Still investigating but nothing till now.
    knowledge is Power , Spread it.
    www.e-tutankhamun.com
    [email protected]
    AIM:AhmedFouad0 , yahooID:xor2004

  3. #3
    Join Date
    Aug 2002
    Location
    Atlanta, GA
    Posts
    1,114
    We had a lot of issues with both versions. It looks like some of the same things we saw.

    We are now using SUSE 9.0 - 64 and it works great. I had not used SUSE before and it still takes me a little time to find some things but it has been rock solid for about 4 monhts.

    Running on Dual AMD 246 with 2 Gb ram and 4x160 SATA raid drives using a 3ware raid card.
    SiteSouth
    Atlanta, GA and Las Vegas, NV. Colocation

Posting Permissions

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