Subject: Re: [sv-ac] alternative proposal to "syn1": remove binary delay
From: Adam Krolnik (krolnik@lsil.com)
Date: Fri Feb 14 2003 - 11:05:02 PST
stra
Hi John,
In another mail, you wrote:
"I think the perceived difference is coming from ";" being
a non-overlapping operator and "=>" (currently defined) an overlapping operator."
I think this is why the forms I wrote earlier look different.
Going with your and Cindy's recommendation of two implication operators, the forms
from my previous mail:
1. (A; [2] b) ok
2. (A; [1:2] b) ok
3. sequence seqC = ([1:10] c);
(A; [0] seq) ok.
4. (A => B); ok
5. (A => 1; B) ok but *1.
6. (A => 1; [1:10] B) ok but *1
7. (A => [1:11] B) ok *2
Written with the nonoverlapping operator would match more closely with
the forms that define sequences with a nonoverlapping concatenation operation.
Note: read these as if => is non-overlapping (though currently its not.)
4. (A => [0] B); ok - like #3 where want A && c as matching.
5. (A => B) ok - just like A; B
6. (A => [1:10] B) ok - numbering matches "1 to 10 cycles later."
7. (A => [0:10] B) ok - like #6, but allowing overlap of A & B.
Thus (nonoverlapping versions)
cover (A; [1:4] B)
assert (A => [1:4] B)
Would imply the same sequences. And if you had overlapping versions
cover (A # [1:4] B)
assert (A &> [1:4] B)
Would also imply the same sequence.
Adam Krolnik
Verification Mgr.
LSI Logic Corp.
Plano TX. 75074
This archive was generated by hypermail 2b28 : Fri Feb 14 2003 - 11:07:48 PST