Untitled

The talk yesterday evening was pretty interesting, although it seemed to be a bit promotional. Jim Mitchell is working on creating a new architecture for high performance computers. some of the ideas are really novel and innovative. others will be difficult to make viable. One of the neatest things is that they’re using proximity to transmit signals between processors. That is, there is an array of capacitors where half is on one chip and the other half is on an adjacent chip, using this way of communicating allows them to get over the pin density problem and they can put these contacts on the chip with lithography techniques. It gives them a way to create very high bandwidth communication between units. The way they’re planning on laying this thing out is interesting. by putting the proximity transmitters on 4 corners of every chip they can lay out a series of chips in a grid 2 levels high. Effectively this gets rid of a number of problems.. the bus speeds and memory access speeds should be amazing. but the heat generated by one collection of these units would be enormous, and thus it is necessary to spray the chips with water from the bottom to keep the system cool.

Some interesting insights.

Today, I’m in school working on my graphics homework. This assignment is not a very interesting one. But I got 10/10 on my tetris game assignment.

I also did a bit of searching for jobs. I’m at the point where I’ve given up on the regular resources, and I’m going to start flooding companies with my resume. Hopefully I’ll find the time to do that this week.

Also, I want to start contributing to some open-source projects, just little patches. To get me started. The problem is that I’ll need to get the newest cvs version of all the stuff I want to work on. And that will take a while for me to get around to do.