Winner of Name that Ware August 2005!

This past month’s name that ware was again a breeze for the contestants. I may take Nate’s suggestion for not this month’s, but for the month after’s contest (blocking out major manufacturer labels and part numbers to try and up the ante). For now, I’m just glad I found the time to put up a new ware this month and declare a winner for last month.

DavidR had the first correct response with his clever SHA1 sum hashed response; congratulations to DavidR! The board is indeed an Intel Paragon node board with an I/O daughterboard. Drop me a line for your prize. Nate as always had impressive comments and impressive response time. I’ve gotta go dig around my archives of obscure boards for more challenging stuff.

A little short on time, I’m not going to post the detailed breakdown of the board’s contents, as DavidR and Nate already did a great job of that in their responses. As for the fatal flaw, back at MIT my PhD advisor, Tom Knight, handed me one of these boards and asked me, “what’s wrong with this board?” and my response–which we can debate on its correctness–was “the tiny network connector on the left.” TK did happen to agree with me on this point, as this machine was bandwidth-starved and had trouble getting scalable performance on programs that did not partition “just so” to the machine’s granularity and architecture. In contrast, some of the better balanced machines out there, such as the Cray or IBM supercomputers, seem to be made out of mostly wires and connectors with a few chips interspersed throughout. It’s all about the wires, baby!

Comments are closed.