Reading Max Maxfield's article on the way thing were in Chip Design magazine made me want to share my story... I started in the semiconductor industry in 1984 when 3 micron technology was at it's peak. My first job at NCR Microelectronics in Ft. Collins Colorado was to measure metal and poly lines with a scale and determine the RC loads for critical nets. Any poly line over 1000 microns long was to be reviewed.
I eventually worked my way up the ladder and became a place and route guy using a public domain router called MP2D which ran on a DEC /VT100. These routes would sometimes run for days - then crash. We never understood exactly how this crash happened but did realise if we rotated the chip 90 degrees it would/could complete the route.
We didn't have a LVS program and instead we would check the "block to block" connections and then run a design rule check. This block to block check didn't look at the transistors inside the block, rather only the connection to each other. I would then wait a day or two for the design rule check to complete only to find that it crashed on occasion also.
That year we complete approximately 10 chips with those room sized computers and a team of engineers. Last year/2008, as a contractor, I completed 10 full chips using Tanner Tools on my PC out of my home office.
Comments (0)
