Subject: assigning 'X'
From: Muzaffer Kal (muzaffer@dspia.com)
Date: Thu Mar 28 2002 - 01:08:28 PST
hi,
I think 1364.1 and the "RTL coding styles that yield simulation and
synthesis mismatches" document are not consistent in this area. 1364.1 says
that "the value x may be used as a primary on the RHS ..." in Section 5.5
"Support for values x and z" and in Section 6.1.1 "Case decoding attributes"
last example it is claimed that assigning X is "much safer" than a full
case. But in Section 5.1 Assigning ‘X’ of the RCSTYS&SM, this is highly
discouraged. I think we need to somehow make these two attitudes consistent.
Also in RCSTYS&SM, section 5.1, it is claimed that code8b "synthesizes to a
smaller and faster implementation than" code8a. I don't see how one can make
this claim. As y is a don't care when s is 2'b11, if assigning c to it is
the fastest solution, I think the synthesizer should be able to figure that
out. Also why should assigning c to y should be faster and smaller than
assigning, say, b ? I think code8a should be at least as fast/small as code
code8b if not better.
regards,
Muzaffer Kal 408.654.9573
DSPIA Inc.
http://www.dspia.com
ASIC/FPGA design/verification consulting specializing in DSP algorithm
implementations
This archive was generated by hypermail 2b28 : Thu Mar 28 2002 - 01:11:14 PST