• Join Us |
  • |
  • Sign in with:

Shacknews & AMD

by Steve Gibson, Mar 13, 2002 4:37am PST
Related Topics – Shacknews

Well, as a few people have noticed Shacknews is now powered by AMD. I managed to convince them that all 10 of you guys who visit the page were worth throwing some hardware at. It's a very simple deal, they give us hardware, we give them a powered by logo on the website. No strings attached, and the servers are ours to keep no matter what is said here. What kind of server are we running? Well actually we're running two. :)

SERVER #1 ·    DUAL AMD Athlon MP(tm) 2000+ processors ·    AMD 760-MP(tm) Chipset based motherboard ·    2 SCSI Controllers ·    6 18GB 15K SCSI Hard Drives ·    3GB DDR266 REGISTERED MEMORY SERVER #2 ·    DUAL AMD Athlon MP(tm) 2000+ processors ·    AMD 760-MP(tm) Chipset based motherboard ·    2 SCSI Controllers ·    8 18GB 15K SCSI Hard Drives ·    2GB DDR266 REGISTERED MEMORY
There are some pictures, load average charts, and more detailed specs here. The two servers are hosted in Dallas, Texas courtesy of another one of our sponsors, Speakeasy.net. How in the world are we managing to get involved with such large companies? Absolute dumb luck!


Read all the News


Comments

« Previous 1 2 Next »
64 Threads* | 146 Comments
  • #71
    Refresh Thread
    By: Ardvaark x
    Reply
    <geek-mode>
    Because I'm a unix geek, here's an explanation of the load averages:

    For the non-unix geeks. A load of 2 on a dual cpu system generally means the system is maxed out and cant process any faster. A load of 25 means all hell has broken loose.

    While there is some disagreement about the actual parameters used to calculate the load averages, especially between differing flavors of unix, the usual calculations most often involve the number of processes queued up to be run over a period of time.

    The man page for the /proc psuedo-filesystem says:

    loadavg
    The load average numbers give the number of jobs in
    the run queue averaged over 1, 5 and 15 minutes.

    So, if the load average is 1.0, there is enough load on the CPU so that, on average, there will always be one thing for it to do at the end of a minute.

    In real-life situations, this does not always mean the system is being overwhelmed. A box running a low priority CPU-intensive application, such as SETI or an Unreal server, will always have a load average of at least 1.0, since those types of programs are always in the run queue needing CPU time.

    Getting above that by much can be bad, however, and Steve is right when he says that a load average of 25 indicates that all hell as broken loose. Quite honestly, I'm rather impressed the server stayed up as long as it did under that kind of intense thrashing.
    </geek-mode>
    Mar 13, 2002 8:25am PST

    Thread Truncated. Click to see all 3 replies.