??? 11/13/06 17:30 Read: times |
#127889 - senior? Responding to: ???'s previous message |
Richard Erlacher said:
I'm not expert in VHDL and I'm completely illiterate in Verilog. I'm not as experienced with FPGA implementation as you either. I realize this, and what chaps my hide (and the hides of others, I'm sure) is that someone who's admittedly a non-expert is giving others advice. Keep in mind that senior engineers went to school before there was 'C' and before there was VHISIC and VHDL. So, after doing this for almost 20 years, I suppose I'm not a "senior" engineer, in spite of what my business card might say? While I like FPGA and CPLD application, I don't like the way in which vendors present them to the user community. I once lived on "evaluation boards" which I could string together in order to build large and complex functional systems in order to prove concepts, and, afterward, tear them down and put them back on the shelf and use them again later. Todays eval boards are like 10 pounds of sh*t in a 5-pound bag, with the "features" always in the way. If it weren't for the short market life of current-generation programmable logic, I'd just build my own boards. And as someone who actually does FPGAs for a living, I've never used an eval board as the basis for anything. It's easier to put the necessary hardware (like fast ADCs and support analog) and interface stuff on a PCB and get a board built. We are using an eval board for a Virtex-4 w/PPC design, but that's mainly to get a head start on the PPC software development. Guys who haven't been doing programmable logic have been drawing schematics with discrete (MSI/SSI) logic and MCU's. For them to abandon what they've been doing and start over, cold, in an unfamiliar HDL is too much "future shock" for them, IMHO, so that's why I've promoted sticking with schematic entry. ... and using 74xxx-equivalent logic symbols is remarkably inefficient. One of the beauties of an HDL is that you describe the logic YOU WANT, without having to see which MSI device is the closest fit. And you haven't answered ANY of my questions or concerns, ESPECIALLY regarding static timing analysis. -a |