31 Mar
2016
31 Mar
'16
6:03 p.m.
On Mar 31, 2016, at 4:39 PM, ben <bfranchuk at
jetnet.ab.ca> wrote:
On 3/31/2016 1:38 PM, ethan at 757.org wrote:
FPGA's can match the TIMING and HARDWARE BUGS better, if given half a chance.
That depends.
If you have a very good functional specification of a processor, you can build a
behavioral model that will execute the same code. But it won't match the bugs (since
those probably aren't in the spec) and it certainly won't match the timing.
You can match timing and bugs, but for that you'd need to have access to the original
design, at the gate level or close to it. That is occasionally possible (for example, the
CDC 6600 design data exists to that level of detail). But it is quite rare to find that.
In fact, even a functional spec good enough to do an FPGA behavioral model is rare. Most
of the time you find yourself reverse engineering things from vague descriptions, just as
emulator software programmers do. The PDP-11 FPGA designs I have seen are examples of
this.
paul
A
raspberry pi 2 fits in an even smaller box, and is a lot cheaper...
FPGA can be configured to match the original hardware. Pis run Linux and
then app on it which makes performence unstable?