1 Feb
1998
1 Feb
'98
2:19 p.m.
Yes, but I've never met an FPGA (and I've used
most Xilinx families,
Altera, etc parts) where you could reassign power pins. Also, most (all?)
have a high-speed clock line that you should use for your master clock if
at all possible. FPGAs are not a 'drop-in' replacement for just any logic
chip.
I would like to make it clear that I am talking about possibilties ten or
twenty years from now. I realize that today's FPGAs would not stand a
chance replacing anything better than something more complex as a 2901.
Impossible. There could be hidden registers that were
never accessed
while you're looking at it. You would have to check _every_ possible mode
and _know_ you'd checked every possible mode. That's impossible if you
don't even have a specification of the chip.
Maybe not 100 percent, but maybe really close. Once again, think when you
will have something the power of a big T3E on your desk.
That's not the worry, although fitting that sort
of logic into _any_ FPGA
would be an interesting task.
Gate arrays and FPGAs are not THAT different. Anyway, real gate arrays are
getting like circuit boards - quick turnaround for just about anyone.
Someday they might even be so reasonably priced that any hobbiest could
special order one, just like the small circuit board shops.
Sure, and standard macros exist, and will continue to
exist, for most
FPGA devices. THe problem is that a chip this complex is going to have
some undocumented features that may or may not be used. Finding them,
figuring out how to implement them, etc, is going to be a lot of work.
If you do not get it correct the first time, well, then you get to have
some fun.
William Donzelli
william(a)ans.net