Re: Getting on the Spartan3e carry chain.

Symon wrote:
Hi John,
Thanks for your reply! You're correct it's the Synplify synthesis tool that
yields the 'bad' mode with tbxcy. Remarkably, it would appear that ISE8.2
was clever enough to change the EDIF circuit to give better timing results,
but ISE10.1 doesn't bother. So, I'll fix the source code. Thanks for your
suggestions, I'll try them out. Maybe there are some synthesis options to
try also.
I wonder, did you ever try instantiating the first MUXCY of the chain to
force the correct behaviour? Gives, horrible looking source code though, I
suppose!
Cheers, Syms.

I've used the instantiated MUXCY trick before; I hate to do it.
You're right that it's not too pretty. I can usually push the rope
pretty well with Synplify (Verilog) so I have strong hopes for your
success.

- John_H
.