Slashdot Mirror

← Back to Stories (view on slashdot.org)

Zalman Showcase Massive P4 Heatsink

Posted by CmdrTaco on from the gonna-need-a-bigger-case dept.

Kez writes "I couldn't express the size of this heatsink in the space provided for the subject of this post. It's the size of a small country and when the fan is running, turbulence from it means a no-fly zone needs to enforced above it. At Hexus.net we've got a picture of this behemoth."

3 of 396 comments (clear)

  1. Re:Why oh why, slashdotted before the first commen by numbski · · Score: 4, Interesting

    I've been passing this around for a while, trying to get it noticed without appearing to be a Karma whore:

    How to fix the slashdot effect.

    Any mod points would be hot. :) I don't need the mod points, but I think this idea is important enough to be noticed, and this is really the only forum for which to discuss it. :\

    --

    Karma: Chameleon (mostly due to the fact that you come and go).

  2. Definitely not the biggest by Anonymous Coward · · Score: 1, Interesting

    This guy obviously didn't see the pictures of the radiator from a Toyota Camry connected to heat pipes on a CPU (sorry, too lazy to look for link). Now THAT was big. :)

  3. Re:Why oh why, slashdotted before the first commen by jdion · · Score: 2, Interesting

    I agree with the above statement,

    But twenty years ago... how much bandwidth was avaliable to the general public?

    I'll be glad to stick a PII 350 running a barebones *NIX webserver with Java any day, but only if my pipe is any less than a OC-1.

    Let's just think about how many people would be required to flood such a link.....

    I'll be generous here, and say that most people using broadband have about a 2Mbps connection. A normal sized webpage usually runs about 30-75kb (including advertising and thumbnails).

    The page is probably cached in RAM, so I'm ignoring hard drive bottlenecks. Even at 75kbs, a 2Mbps connection can pull that page in about 3.5 seconds. Thanks to the checks and balances of TCP, an ACK packet must be resent back to the server. (This is not a knock, just a comment). Going back to the example: An OC-1 connection is quickly flooded by 26 people connecting at the same time. I'm pretty sure this happens! 3.5 seconds later... everybody has their webpage.

    Now in the real world, anywhere between 3k and 5k connect to a slashdot page at any given time. Granted, not all 5k connect at the same time, but I would venture to say that about 300 people connect in the same 3 second window. That right there is about 600Mbps downspeed. Keep that flood going, and you'll be lucky if the first 50 people get their ACK packets back to the server to keep feeding them data.

    Is TCP poorly designed. Largely, no. Possible solutions: UDP! Someone come up with a UDP standard for page viewing. Is this going to happen in the real world? No... probably not. I would hate to see some website improperly read because a packet didn't reach the destination and so there are holes in the transmission. How about smaller webpages? Five years ago... pages were not nearly as graphic heavy as they are now. Is it required? No. Will it go away. No.

    Simple answer: Slashdot should have a rotation script that displays the link on a series of user accounts after a time period. Why not? They do it with paid users... why not just implement it into a time script. Would this take more processing power from Slashdot... you bet it would, but how much more? Think about how many websites might actually be able to be read by users...

    Wait... do Slashdot users actually RFA?

    Slashdot is a central hub for DDoS attacks. Forget Zombie networks.... post that page on Slashdot... and it's going down.

    Just food for thought.