My primary interest in computing is computer architecture — particularly the design of microprocessors. My posts on several Stack Exchange sites (Electrical Engineering, Computer Science, Super User, and StackOverflow) provide some evidence of my interest. I have also posted substantially on the comp.arch newsgroup and the Real World Technologies Forum.
Although I have never advanced the state of the art or contributed to an actual processor design, one of my ideas was actually implemented by IBM. The exploiting of register file replication in a clustered microarchitecture was used in POWER7 to support four threads at the storage cost of two threads. I had posted this idea on comp.arch in 2010 (27 May; Message-ID: <email@example.com>), but people at IBM must have thought of it at least four years earlier since POWER7 came out in 2010. I have joked that people steal my good ideas, then go back in time and publish or use them before I discovered the ideas.
Andy Glew's CompArch Wiki
Some years ago, Andy Glew let me contribute to his Computer Architecture Wiki. The original site was abandonned, but a new version is in the process of being set up. Some of the original articles are still available through the Internet Archive.
I did not contribute that much to the wiki, but I am somewhat proud of the article on supporting multiple page sizes in a TLB (which I mostly wrote). I think I was also the primary author for the article on skewed associativity.