Subject: always_comb semantics
From: Karen Pieper (Karen.Pieper@synopsys.com)
Date: Wed Oct 02 2002 - 17:30:33 PDT
Hi, all,
I have to come down on the side of keeping the always_{comb ...}
functionality in the language.
These new types of always blocks clearly have the semantic requirements in
place to force the language used to be synthesizable. This shows up
because of the required checks that there be a single driver for the nets
assigned in the always blocks, and in the sensitivity to the internals of
functions.
One of the goals for SystemVerilog is to create a language where systems
can be built correctly and
more quickly. Requiring these checks in every tool will achieve that end
much better than checking done only by a subset of the tools. Time will
be wasted in getting something through one tool and then having an issue be
found by a later one. To the more that we can eliminate that waste, the
more our users will appreciate what we have achieved.
Karen
This archive was generated by hypermail 2b28 : Wed Oct 02 2002 - 17:32:11 PDT