Announcement

Collapse
No announcement yet.

IO_uring Is Maturing Well On Linux For Faster & More Flexible I/O - Benchmarks On Linux 5.6

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

  • IO_uring Is Maturing Well On Linux For Faster & More Flexible I/O - Benchmarks On Linux 5.6

    Phoronix: IO_uring Is Maturing Well On Linux For Faster & More Flexible I/O - Benchmarks On Linux 5.6

    Since its introduction in Linux 5.1, IO_uring has been coming together quite nicely and getting better with each new kernel release. IO_uring is the effort for delivering faster and more efficient I/O by avoiding excess copies and other efficiency improvements over the existing Linux AIO code. Here are some comparison benchmarks off Linux 5.6 Git...

    http://www.phoronix.com/scan.php?pag...IO-uring-Tests

  • #2
    Wow, the performance improvements look awesome!
    I hope to see KDE Plasma implement this for file copy / move operations.

    Comment


    • #3
      Originally posted by Danny3 View Post
      Wow, the performance improvements look awesome!
      I hope to see KDE Plasma implement this for file copy / move operations.
      That would be cool. Does Gnome support this?

      Comment


      • #4
        Originally posted by Volta View Post

        That would be cool. Does Gnome support this?
        I have no idea, but someone like the Linux Mint's developers will probably add it sometime in the future for Cinnamon.

        Comment


        • #5
          and the performance compared to non AIO? AKA old fashioned blocking i/o?

          Comment


          • #6
            The performance increase looks significant.

            I would like to see it implemented in APT, DNF, yum, GNOME Software, and 7zip.

            Comment


            • #7
              Originally posted by Danny3 View Post
              Wow, the performance improvements look awesome!
              I hope to see KDE Plasma implement this for file copy / move operations.
              I have my doubts about whether this would increase the performance of normal operations such as copying and moving files where you read/write large chunks of sequential data.
              I think this is more suitable to asynchronous disk writes with many smaller non-sequential writes, such as a database where you would want to update different columns and rows.

              Comment


              • #8
                Originally posted by uid313 View Post

                I have my doubts about whether this would increase the performance of normal operations such as copying and moving files where you read/write large chunks of sequential data.
                I think this is more suitable to asynchronous disk writes with many smaller non-sequential writes, such as a database where you would want to update different columns and rows.
                I'm thinking that not all the time I copy or move large files like, ISO files, video or audio files, sometimes I copy or move folders which can contain many small files.
                And besides, I think many other stuff can happen in the same time like running an update or apt-get update or install some program, I bet the browser is also doing some reads and writes.
                Probably the system is also logging something to some files all the times.
                Taking all these combined, I think it be very nice if at least some of the programs can use this high IOPS interface and lighten the load.
                Last edited by Danny3; 18 March 2020, 10:48 AM.

                Comment


                • #9
                  Originally posted by Danny3 View Post

                  I have my doubts about whether this would increase the performance of normal operations such as copying and moving files where you read/write large chunks of sequential data.
                  I think this is more suitable to asynchronous disk writes with many smaller non-sequential writes, such as a database where you would want to update different columns and rows.

                  I'm thinking that not all the time I copy or move large files like, ISO files, video or audio files, sometimes I copy or move folders which can contain many small files.
                  And besides, I think many other stuff can happen in the same time like running an update or apt-get update or install some program, I bet the browser is also doing some reads and writes.
                  Probably the system is also logging something to some files all the times.
                  Taking all these combined, I think it be very nice if at least some of the programs can use this high IOPS interface and lighten the load.
                  Please note IO_URING should not improve performance and will not improve performance, if your drive max out moving files today will max out exactly the same with IO_URING.

                  What IO_URING fix is I/O contention and/or latency when handling multiple request at same time simply because it can handle efficiently and asynchronously every operation independently of each other whereas AIO have to deal with threads, fences, spinlocks, queues, etc. etc .etc. which in modern hardware introduces a lot of unnecessary overhead.

                  So technically speaking if something is faster with IO_URING is not because of performance but because your workload was hitting a bottleneck with AIO and now is not, also this will improve greatly Desktop operations responsiveness(in theory) when you have several intensive I/O operations running in parallel(like copying some files, running some DB on the background, loading some game, installing some updates, etc.)

                  Comment


                  • #10
                    Here why this kernel should be integrated in the next Ubuntu release on which many other distros are besed on.

                    Comment

                    Working...
                    X