Is there any sense in referring to "a Beowulf cluster of Beowulf clusters of X" (X in this case is obviously "XBoxes", but same logic applies for all X)? That is, is, say, an 8x cluster of 8x clusters at best equivalent to a 64x cluster? Or are there situations where having some kind of (at least logical, if not physical) cluster hierarchy is an advantage, e.g. to isolate network traffic for closely related tasks?
Maybe you were just quipping, but on the offchance that it was a serious question... I think the point is that previously known fullerenes (Buckyballs etc) weren't planar, i.e. the molecular bonds meet each other to form a three-dimensional structure, so you can't flatten it into a single sheet.
Maybe they got the direction of causality wrong. More work and less errors = less entropy. Productive workers need to release heat to compensate.
This may have worthwhile applications... you could hire someone smart and productive to sit in your living room doing things as an alternative to central heating. Just give him random tasks and watch the temperature rise. Get him to do your tax while he's there.
Or, to move it down a level of sentience, we all know that a productive and useful computer generates one heckuva lot more heat than a useless one... could this measure not just its potential but its actual use? Does your box get hotter the less typos you make? Is the converse true?
Potential for case cooling: the dumber and more random your moves in [insert FPS here], the less heat your machine has to generate to keep the entropy up.
Slashdot Mirror
Random question.
Is there any sense in referring to "a Beowulf cluster of Beowulf clusters of X" (X in this case is obviously "XBoxes", but same logic applies for all X)? That is, is, say, an 8x cluster of 8x clusters at best equivalent to a 64x cluster? Or are there situations where having some kind of (at least logical, if not physical) cluster hierarchy is an advantage, e.g. to isolate network traffic for closely related tasks?
Maybe you were just quipping, but on the offchance that it was a serious question... I think the point is that previously known fullerenes (Buckyballs etc) weren't planar, i.e. the molecular bonds meet each other to form a three-dimensional structure, so you can't flatten it into a single sheet.
Maybe they got the direction of causality wrong. More work and less errors = less entropy. Productive workers need to release heat to compensate.
This may have worthwhile applications... you could hire someone smart and productive to sit in your living room doing things as an alternative to central heating. Just give him random tasks and watch the temperature rise. Get him to do your tax while he's there.
Or, to move it down a level of sentience, we all know that a productive and useful computer generates one heckuva lot more heat than a useless one... could this measure not just its potential but its actual use? Does your box get hotter the less typos you make? Is the converse true?
Potential for case cooling: the dumber and more random your moves in [insert FPS here], the less heat your machine has to generate to keep the entropy up.
Interesting...