Netgate Discussion Forum
    • Categories
    • Recent
    • Tags
    • Popular
    • Users
    • Search
    • Register
    • Login

    Different performance with different versions of pfsense

    Scheduled Pinned Locked Moved General pfSense Questions
    11 Posts 3 Posters 2.8k Views
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • T
      tigs
      last edited by

      I have been playing with an older jetway Atom N550 board to have a sense of what pfsense is like. I did some openssl tests. Here are results:

      blowfish

      pfsense 2.2.5 64 bit: 25138.59k
      pfsense 2.1.5 64 bit: 37027.10k
      pfsense 2.1.5 32 bit: 52074.47k
      openwrt            39400.79k
      ddwrt                      32732.62k

      aes-256 cbc

      pfsense 2.2.5 64 bit: 16659.80k
      pfsense 2.1.5 64 bit: 17016.63k
      pfsense 2.1.5 32 bit: 15216.52k
      openwrt            18658.65k
      ddwrt                      24939.18k

      The best performance is with version 2.1.5 32 bit. To my surprise, the latest version performed very poorly.

      1 Reply Last reply Reply Quote 0
      • T
        tigs
        last edited by

        I have just finished setting up my pfsense with supermicro C2758 mini ITX box with 8G of RAM. Install on a HDD from USB drive. I am kind of disappointed to see its openssl benchmark score is barely better than my 2nd gen Atom N550. Can someone explain why?

        blowfish cbc    72496.47k
        aes-256 cbc    89642.33k

        For reference, here is the score for my Netgear R7000 Running DD-WRT:

        blowfish cbc:  31663.98k

        aes-256 cbc:  27368.36k

        1 Reply Last reply Reply Quote 0
        • H
          heper
          last edited by

          because its still an atom.

          check the aes-gcm performance ;)

          1 Reply Last reply Reply Quote 0
          • T
            tigs
            last edited by

            The netgear r7000 is dual core 1.0Ghz ARM router with no hardware encryption acceleration of any form. This one is a 8-core Rangely 2.4GHz CPU and with AES-NI and quickassist.

            Given the fact that the benchmark score varies with differen version of pfsense, I think it is software issue.
            Pfsense is not optimized.

            1 Reply Last reply Reply Quote 0
            • H
              Harvy66
              last edited by

              Different versions of PFSense have different versions of OpenSSL. Are you sure it isn't a regression in OpenSSL and not PFSense?

              1 Reply Last reply Reply Quote 0
              • H
                heper
                last edited by

                you arent using any AES-NI. and quickassist is not supported (yet) on pfsense.

                1 Reply Last reply Reply Quote 0
                • T
                  tigs
                  last edited by

                  @Harvy66:

                  Different versions of PFSense have different versions of OpenSSL. Are you sure it isn't a regression in OpenSSL and not PFSense?

                  I have no idea about this. I used the public release of pfsense. I did not modify any of the pfsense I have tested. Everything is stock package. After fresh install and enabling the SSH, I tested them via ssh.

                  I expected a lot more from this board based on the hardware spec. I wonder whether people who have the netgate routers can give it a test with their routers.

                  1 Reply Last reply Reply Quote 0
                  • T
                    tigs
                    last edited by

                    I have just tested pfsense openvpn performance with this setup. I have a cable internet of 100M.

                    With version 2.1.5 32 bit: I get 120M down load
                    with version 2.2.5 64 bit, I get only 50-65M, only half of that with older.

                    Without vpn, I get 120-130M.

                    I still believe it is the software optimization issue.

                    1 Reply Last reply Reply Quote 0
                    • H
                      heper
                      last edited by

                      did you try different powerD settings?

                      1 Reply Last reply Reply Quote 0
                      • T
                        tigs
                        last edited by

                        @heper:

                        did you try different powerD settings?

                        Sorry I am still learning. Woul dyou be able to direct me to the instructions? I will test it. Attached is the openvpn result with version 2.1.5 32 bit. It is virtual the same as without vpn.

                        http://www.speedtest.net/my-result/4921681179

                        1 Reply Last reply Reply Quote 0
                        • H
                          heper
                          last edited by

                          System: Advanced: Miscellaneous: PowerD

                          1 Reply Last reply Reply Quote 0
                          • First post
                            Last post
                          Copyright 2025 Rubicon Communications LLC (Netgate). All rights reserved.