Gigabit Ethernet, and Linux -- first observations - SSH

This is a discussion on Gigabit Ethernet, and Linux -- first observations - SSH ; On 2007-09-13, Darren Tucker wrote: > Try it by all means, but don't be surprised if it makes little difference > on a LAN. Oh, one thing that will name a difference: using the patch on the HPN page that ...

+ Reply to Thread
Page 2 of 2 FirstFirst 1 2
Results 21 to 24 of 24

Thread: Gigabit Ethernet, and Linux -- first observations

  1. Re: Gigabit Ethernet, and Linux -- first observations

    On 2007-09-13, Darren Tucker wrote:
    > Try it by all means, but don't be surprised if it makes little difference
    > on a LAN.


    Oh, one thing that will name a difference: using the patch on the HPN
    page that allows the "none" cipher. It will reduce the CPU usage, and
    that might or might not make a difference to the throughput depending
    on what the limiting factor is at that point.

    Some symmetric ciphers are surprisingly cheap on modern hardware.
    My four year old laptop (1.3GHz Pentium M) reports ~186 MByte/s for
    arcfour (via "openssl speed rc4").

    --
    Darren Tucker (dtucker at zip.com.au)
    GPG key 8FF4FA69 / D9A3 86E9 7EEE AF4B B2D4 37C9 C982 80C7 8FF4 FA69
    Good judgement comes with experience. Unfortunately, the experience
    usually comes from bad judgement.

  2. Re: Gigabit Ethernet, and Linux -- first observations

    On Thu, 13 Sep 2007 11:15:38 -0500, Ignoramus32529 wrote:

    > On Thu, 13 Sep 2007 19:07:39 +1000, Darren Tucker
    > wrote:
    >> On 2007-09-13, Darren Tucker wrote:
    >>> Try it by all means, but don't be surprised if it makes little
    >>> difference on a LAN.

    >>
    >> Oh, one thing that will name a difference: using the patch on the HPN
    >> page that allows the "none" cipher. It will reduce the CPU usage, and
    >> that might or might not make a difference to the throughput depending
    >> on what the limiting factor is at that point.
    >>
    >> Some symmetric ciphers are surprisingly cheap on modern hardware. My
    >> four year old laptop (1.3GHz Pentium M) reports ~186 MByte/s for
    >> arcfour (via "openssl speed rc4").
    >>
    >>

    > You know, I can get about 50 MegaBYTES per second using arcfour128.


    Or about 5 times as much, if running on IA64.

  3. Re: Gigabit Ethernet, and Linux -- first observations

    On Thu, 13 Sep 2007 17:04:10 GMT, Ivar Rosquist wrote:
    > On Thu, 13 Sep 2007 11:15:38 -0500, Ignoramus32529 wrote:
    >
    >> On Thu, 13 Sep 2007 19:07:39 +1000, Darren Tucker
    >> wrote:
    >>> On 2007-09-13, Darren Tucker wrote:
    >>>> Try it by all means, but don't be surprised if it makes little
    >>>> difference on a LAN.
    >>>
    >>> Oh, one thing that will name a difference: using the patch on the HPN
    >>> page that allows the "none" cipher. It will reduce the CPU usage, and
    >>> that might or might not make a difference to the throughput depending
    >>> on what the limiting factor is at that point.
    >>>
    >>> Some symmetric ciphers are surprisingly cheap on modern hardware. My
    >>> four year old laptop (1.3GHz Pentium M) reports ~186 MByte/s for
    >>> arcfour (via "openssl speed rc4").
    >>>
    >>>

    >> You know, I can get about 50 MegaBYTES per second using arcfour128.

    >
    > Or about 5 times as much, if running on IA64.


    You mean a 64 bit CPU? Sorry if I am missing something obvious.

    My laptop is 64 bit, the rest is, unfortunately, 32 bit.

    i

  4. Re: Gigabit Ethernet, and Linux -- first observations

    On Thu, 13 Sep 2007 14:59:20 -0500, Ignoramus32529 wrote:

    > On Thu, 13 Sep 2007 17:04:10 GMT, Ivar Rosquist
    > wrote:
    >> On Thu, 13 Sep 2007 11:15:38 -0500, Ignoramus32529 wrote:
    >>
    >>> On Thu, 13 Sep 2007 19:07:39 +1000, Darren Tucker
    >>> wrote:
    >>>> On 2007-09-13, Darren Tucker wrote:
    >>>>> Try it by all means, but don't be surprised if it makes little
    >>>>> difference on a LAN.
    >>>>
    >>>> Oh, one thing that will name a difference: using the patch on the HPN
    >>>> page that allows the "none" cipher. It will reduce the CPU usage,
    >>>> and that might or might not make a difference to the throughput
    >>>> depending on what the limiting factor is at that point.
    >>>>
    >>>> Some symmetric ciphers are surprisingly cheap on modern hardware. My
    >>>> four year old laptop (1.3GHz Pentium M) reports ~186 MByte/s for
    >>>> arcfour (via "openssl speed rc4").
    >>>>
    >>>>
    >>> You know, I can get about 50 MegaBYTES per second using arcfour128.

    >>
    >> Or about 5 times as much, if running on IA64.

    >
    > You mean a 64 bit CPU? Sorry if I am missing something obvious.


    I mean the Itanium 2 processor.

    >
    > My laptop is 64 bit, the rest is, unfortunately, 32 bit.
    >
    > i



+ Reply to Thread
Page 2 of 2 FirstFirst 1 2