Announcement

Collapse
No announcement yet.

Ubuntu 8.04 LTS Through 10.10 Virtualization Benchmarks

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Ubuntu 8.04 LTS Through 10.10 Virtualization Benchmarks

    Phoronix: Ubuntu 8.04 LTS Through 10.10 Virtualization Benchmarks

    Earlier this month we delivered Ubuntu 10.10 benchmarks from some different hardware comparing the performance of this "Maverick Meerkat" release to that of Ubuntu 9.10 and 10.04.1 LTS. The results were interesting, but since then we have had the time to complete additional tests. In this benchmarking roundabout, we decided to see how the performance of every release from Ubuntu 8.04 LTS through the new Ubuntu Linux release performs when tested in a virtualized environment using Linux's KVM virtualization. Here are the virtualized guest results for Ubuntu 8.04.4 LTS, 8.10, 9.04, 9.10, 10.04.1 LTS, and 10.10.

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2


    Same old whining about the ext4... Would it be so hard to test some different filesystems like ext3, xfs, btrfs?!

    Comment


    • #3
      Originally posted by Ratbert View Post


      Same old whining about the ext4... Would it be so hard to test some different filesystems like ext3, xfs, btrfs?!
      These tests are designed to look at the out-of-the-box performance of Ubuntu, if you begin changing out the file-system and other components, you're no longer providing a stock look at the operating system.

      In articles where the focus is on the file-system, we of course test different file-systems, but not if trying to look at the performance of Ubuntu over time.
      Michael Larabel
      https://www.michaellarabel.com/

      Comment


      • #4
        still, it would be really interesting to see the same benchmarks with 10.04 on ext3 and on ext4, just to see if the major difference was due to the filesystem or to other changes in ubuntu/the kernel.

        Comment


        • #5
          comparison back to native would be nice

          ... as well as VirtualBox OSE. I've never seen any stats, but I'd be surprised if there were more KVM installs than virtualbox on the desktop anyway.

          Comment


          • #6
            Originally posted by jure1873 View Post
            still, it would be really interesting to see the same benchmarks with 10.04 on ext3 and on ext4, just to see if the major difference was due to the filesystem or to other changes in ubuntu/the kernel.
            As Michael always offers, once you move out from the default install the options become extreme and can be configured in any particular way someone would want.

            If you are willing to document (wiki, email, etc) your suggested test configuration, then Michael is generally willing to do a follow up. To date, I don't believe anyone has actually written any suggested configurations.

            Comment


            • #7
              Ext4 with PostgreSQL

              For PostgreSQL pgbench 8.4.0

              Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

              The PostgreSQL performance is a similar story to that of our Apache test profile where the transactions per second improved greatly during the switch to EXT4, but with PostgreSQL the numbers continued to improve between 9.10 and 10.04.1 ....

              Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

              The PostgreSQL database server takes a large performance hit between Ubuntu 9.10 and 10.04.1 LTS, which is due to EXT4 changes in the Linux 2.6.32 kernel as we confirmed previously.
              I found that with its result these two statements are against each other. Can anyone explain why?

              Comment


              • #8
                Originally posted by phoronix View Post
                Phoronix: Ubuntu 8.04 LTS Through 10.10 Virtualization Benchmarks

                Earlier this month we delivered Ubuntu 10.10 benchmarks from some different hardware comparing the performance of this "Maverick Meerkat" release to that of Ubuntu 9.10 and 10.04.1 LTS. The results were interesting, but since then we have had the time to complete additional tests. In this benchmarking roundabout, we decided to see how the performance of every release from Ubuntu 8.04 LTS through the new Ubuntu Linux release performs when tested in a virtualized environment using Linux's KVM virtualization. Here are the virtualized guest results for Ubuntu 8.04.4 LTS, 8.10, 9.04, 9.10, 10.04.1 LTS, and 10.10.

                http://www.phoronix.com/vr.php?view=15380
                That was a very interesting article!! Thank you for the info!
                I'm really interested in VM related benchmarks and would like to see more articles like this in the future!
                Also it would be so nice of you if you could also include the benchmarks from the actual host to see how they compare to the guests

                Comment


                • #9
                  Originally posted by waveyus View Post
                  For PostgreSQL pgbench 8.4.0

                  I found that with its result these two statements are against each other. Can anyone explain why?
                  The software subsystems under test are different. Well, actually the VM based testing is a superset.

                  For the native testing, IO operations are a lot simpler and in the case of ext4, barriers are used to ensure filesystem consistency. This penalizes numbers where fsyncs and the like are used (since it flushes data to the disk).

                  When you move to a VM based environment the guest goes through the guest filesystem, the guest block device drivers, the host qemu block device emulation and into the host filesystem, host disk drivers and then into the disk. Barriers are a filesystem construct that forces some level consistency in the data written to disks.

                  Adding that complexity results in sometimes vastly different behaviours, particularly when it comes to barriers. With that many layers of sw, there are many options about how you can deal with the barriers. For example, six or so months ago I had to dig through KVM/QEMU/SQLITE/Ubuntu for a similar issue - barriers were being ignored. You can then go to barriers being simplistically honored, or you can look to batch the collection of barriers from the upper layers, etc.

                  Michael's statement for virtualization based tests is that they had a similar change in performance to Apache. He didn't compare the KVM guest behaviour to the native behaviour differences.

                  Comment


                  • #10
                    Originally posted by mtippett View Post
                    The software subsystems under test are different. Well, actually the VM based testing is a superset.

                    For the native testing, IO operations are a lot simpler and in the case of ext4, barriers are used to ensure filesystem consistency. This penalizes numbers where fsyncs and the like are used (since it flushes data to the disk).

                    When you move to a VM based environment the guest goes through the guest filesystem, the guest block device drivers, the host qemu block device emulation and into the host filesystem, host disk drivers and then into the disk. Barriers are a filesystem construct that forces some level consistency in the data written to disks.

                    Adding that complexity results in sometimes vastly different behaviours, particularly when it comes to barriers. With that many layers of sw, there are many options about how you can deal with the barriers. For example, six or so months ago I had to dig through KVM/QEMU/SQLITE/Ubuntu for a similar issue - barriers were being ignored. You can then go to barriers being simplistically honored, or you can look to batch the collection of barriers from the upper layers, etc.

                    Michael's statement for virtualization based tests is that they had a similar change in performance to Apache. He didn't compare the KVM guest behaviour to the native behaviour differences.

                    Thanks.
                    I noticed one difference. The later VM based testing used 10.04.1 (August 2010)release build while previous testings used 10.04 (April 2010).

                    Is it worth having a comparison test between these two in the near future?

                    Comment

                    Working...
                    X