Announcement

Collapse
No announcement yet.

Linux 2.6.38 EXT4, Btrfs File-System Benchmarks

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

  • Linux 2.6.38 EXT4, Btrfs File-System Benchmarks

    Phoronix: Linux 2.6.38 EXT4, Btrfs File-System Benchmarks

    Along with finally delivering Intel Gallium3D driver benchmarks comparing this unofficial, proof-of-concept i915/945 Gallium3D driver to Intel's official classic Mesa driver, there's also our benchmarks of the EXT4 and Btrfs file-systems from the Linux 2.6.38 kernel. These exclusive tests are coming this weekend as part of OpenBenchmarking.org being publicly available for the first day...

    http://www.phoronix.com/vr.php?view=OTEzOQ

  • #2
    I'd really appreciate it if you could use old-fashioned mechanical disks for benchmarking at least half of the time. Considerably more people have them than SSDs, and the performance profiles are vastly different -- results measured with one have almost no relevance to the other. Thanks.

    Comment


    • #3
      reiser4?

      As usual, reiser4 is missing.

      Saddens me.

      Comment


      • #4
        didnt past benchmarks show huge improvments in btfs performonce when compression was swched on?

        Comment


        • #5
          Originally posted by ayumu View Post
          As usual, reiser4 is missing.

          Saddens me.
          Reiser4 is DOA.

          Comment


          • #6
            Originally posted by drag View Post
            Reiser4 is DOA.
            • Its current website is: https://reiser4.wiki.kernel.org/index.php/Main_Page
            • It's been ready for years.
            • It's fast: https://reiser4.wiki.kernel.org/index.php/Benchmarks
            • I've used it for years and it gives me awesome performance.
            • It's got legendary reliability: I've never lost a byte to it. I was even able to recover a lot of data from a ddrescued HD which failed pretty heavily on the hardware side of things, with only a few portions of the disk being readable.
            • It's on -mm, pending only on porting it to some non-reiser4-related-but-useful Linux vfs improvements which afaik aren't finished yet.
            How is it DOA? Care to explain?

            Comment


            • #7
              It's got legendary reliability: I've never lost a byte to it. I was even able to recover a lot of data from a ddrescued HD which failed pretty heavily on the hardware side of things, with only a few portions of the disk being readable.
              I've never lost a byte to Linux's vfat either and it's extremely fast. Does that mean it's as good as Reiserfs?

              It's on -mm, pending only on porting it to some non-reiser4-related-but-useful Linux vfs improvements which afaik aren't finished yet.
              How is it DOA? Care to explain?
              Because it's been under development since 2004 and is still in -mm tree, while a much sophisticated file system exists in the mainline kernel with far more development resources and organizations backing it.

              Comment


              • #8
                I'm not aware of any such filesystem.

                Oh. I hope you don't mean ext4 or btrfs. You can't be serious.

                Originally posted by drag View Post
                I've never lost a byte to Linux's vfat either and it's extremely fast. Does that mean it's as good as Reiserfs?



                Because it's been under development since 2004 and is still in -mm tree, while a much sophisticated file system exists in the mainline kernel with far more development resources and organizations backing it.

                Comment


                • #9
                  Originally posted by drag View Post
                  I've never lost a byte to Linux's vfat either and it's extremely fast. Does that mean it's as good as Reiserfs?
                  CS101 - p->q doesn't mean q->p....:-D

                  At least know basic computer science when you talk about computers.

                  Comment


                  • #10
                    Originally posted by devsk View Post
                    CS101 - p->q doesn't mean q->p....:-D

                    At least know basic computer science when you talk about computers.
                    It doesn't matter which way you read it; neither one usefully substantiates the reliability of a filesystem.

                    Comment


                    • #11
                      Originally posted by devsk View Post
                      CS101 - p->q doesn't mean q->p....:-D
                      Its a flaw in classic logic related, invented by humans, its a flaw that was missused by sophists many times. As well as sharp 1/0 without third state or analogous variability, ie you either breath in or breath out, nothing inbetween.

                      Reiser4 was clearly heavily sabotaged, as it would bring huge advantage to linux. Hans has made two mistakes - married an opportunistic woman and lost the situation out of his control. I heard the system is usable, but it is lobbied by RedHat for whatever reason, and only russian developer,himself not very much interested, patches it for every new kernel. The only problems I heard were loosing data integrity(? don't know if its resolved) and mounting the system with image of this or other reiser4 system residing on that disk in form of file. The system itself should be (much) more advanced than btrfs itself.

                      Very sad indeed.

                      Comment

                      Working...
                      X