Talk:System X (supercomputer)

Was this simply a stunt, or was any actual science done on this "supercomputer"? It seems they just ran the Linpack benchmark, and then tore it apart. Westwind273 22:47, 17 January 2007 (UTC)

They tried to prove you could put together a supercomputer from commodity parts. Unfortunately, they didn't widely report that the consumer-grade version was a failure, because of cosmic rays flipping bits. -covracer (talk) 02:08, 31 January 2010 (UTC)

Errors in Computation on the Original System X

I go to school at Tech and learned this information from someone who worked on the project: The first attempt at putting together the supercomputer used consumer-grade parts and couldn't return correct answers because of cosmic rays flipping bits in its RAM. Unfortunately, the only documentation I can find of this is on the Apple profiles page documenting the upgrade to server-grade components [1] that included error-correction in the memory controller:

Error correction is vitally important for the kinds of “grand-challenge” computational science problems handled by System X. With these problems, says Dr. Varadarajan, “nobody knows what the data is supposed to look like. Without a way to correct for errors, scientists have to repeat a run five or six times. And, even then, all of their results may only point roughly in the same direction. Their machines are telling them nothing useful.”

This shortcoming of the original system potentially explains the otherwise mysteriously small amount of information online (and oddly worded official descriptions) regarding the original System X. Quietly upgrading to different components and pretending the thing worked the whole time seems like academic dishonesty to me, but I've tried to document only the bits of the situation that are verifiable and will write all original research on another forum. -covracer (talk) 02:52, 31 January 2010 (UTC)