Page 2 of 2 FirstFirst 1 2
Results 11 to 11 of 11
  1. #11
    ROG Guru: Green Belt Array fostert's Avatar
    Join Date
    Oct 2011
    Reputation
    54
    Posts
    706

    Quote Originally Posted by fostert View Post
    Agreed the disk speed WEI number is likely inflated at least in part by caching, but I believe there is probably a real performance increase there too because of the journaling filesystem performance advantages versus NTFS. But I am curious so tomorrow I'll run the WEI again in the VM after I disable Linux' caching. I honestly can't see the RAM increase being due to caching in any way, since there is nowhere to cache beyond RAM (except the CPU caches, which should affect both virtual and native scores evenly), so thats gotta be a 100% real increase.
    Okidokie; after booting Windows 7 and VMWare Workstation 8.0 under Linux, I flushed the disk cache completely (command: sync ; sudo sh -c 'echo 3 > /proc/sys/vm/drop_caches') and then ran the WEI bench in the VM. I again flushed the cache clean and ran the WEI for a second time. There should be no performance advantage from digging into the cache for system files, since it is wiped clean before each run. Scores identical: 7.3,7.9,6.0,6.0,6.8 (recall my native windoiws boot scores are 7.4,7.6,7.2,7.2,5.9)

    So the disk score has not changed and one is tempted tho think that performance of the disk under a VM is substantially higher... however, I did some digging into why. The real reason is that Windows limited ALL physical platter disks artificially to 5.9 UNLESS they are connected to a hardware RAID controller. Even if you have hardware controlled RAID, it is still 5.9 max.
    The VM's disk controller tricks Windows by pretending like its hard drive is an LSI SAS controller, so Windows actually goes ahead and scores it based on transfer speed and seek times without simply limiting it to 5.9. See this thread on overclock.net.

    So, in other words, the WEI Disk score is meaningless if you have a HDD, since Windows will run the tests, determine its a platter type drive and blindly assigns a score of 5.9. This explains something I have long noticed too: that 5400 rpm drives get the same WEI score as 7200 rpm drives!
    The score of 6.8 I am getting in the VM is therefore a "real" score for my drive, but it is not a performance increase relative to the 5.9 since that was not obtained from a real test. The moral of the story: there is no way of knowing with the WEI what you have gained in an upgrade relative to what you had, so its a useless benchmark.
    --
    G74SX-CST1-CBIL, i7 2630QM 2GHz
    32GB DDR3 RAM @1333MHz
    GTX560M 3GB DDR5 (192 bit)
    17.3" LED 1920x1080
    Sentelic TP, BIOS 203
    Debian Linux Wheezy (Testing) Kernel 3.2, NVIDIA 295.40

Page 2 of 2 FirstFirst 1 2

Posting Permissions

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