Hi Gord, Yes! I had argued earlier that we needed some definition of the random behavior of free variables. I thought that, at least in a simulation context, the free variables should be randomly assigned in a similar manner to to rand variables, with repeatability, and a capacity for randomization. The semantics in a formal verification context can remain the same. Tom Gordon Vreugdenhil wrote: > > > Hmm. X doesn't seem to be a good idea to me; that will essentially > make it impossible to satisfy most semi-interesting dependent > freevar assignment assumptions since computations involving X generally > result in X. In other words, won't X results in most cases just end up > causing most assertions to fail? That would seem to be an unfortunate > result > of standardizing on that approach. > > I don't think that the RNG approach is really that hard, is it? > Why not just say that there is one system-wide RNG that is used for > all freevar assignments and that such an RNG is assigned as though > it were in a separate top-level module. That should, I think, > preserve random stability even when the number of checkers > changes within the design. In addition, since RNG results don't > produce X or Z values, I think that the likelihood of having > reasonable results is much higher. > > Gord -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.Received on Fri Mar 21 21:59:35 2008
This archive was generated by hypermail 2.1.8 : Fri Mar 21 2008 - 22:02:37 PDT