From: Chao Wang (Wangc_at_colorado.edu)
Date: Wed Nov 05 2003 - 01:29:39 MST
Yimin,
I am not sure if I understood your question correctly (if I didn't, by all
means send us another email.)
It seems that you want to do the following: For each OR gate, z=x+y, you
want to replace it by
z_prime = x + y
z = z_prime * pi_ORgate_is_normal.
Of course this can be encoded by BLIF. For example,
.names z_prime pi_ORgate_is_normal z
11 1
However, when pi_ORgate_is_normal=0, the output of z won't be an 'X' in
this case, because all the signals are binary.
If you want z to take arbitrary value (either 0 or 1, but is arbitrarily
selected), you can add a psuedo-input variable called rand_choice:
.names rand_choice
0
1
.names z_prime rand_choice pi_ORgate_is_normal z
--1 =z_prime
--0 =rand_choice
Chao
On Tue, 4 Nov 2003, [gb2312] Yimin Zhang wrote:
> Dear Friend:
> I want to make some expriments over ISCAS bechmark,checking whether OBDD can be used as a tools of diagnosis,but I have to modify the standard BLIF file for the purpose of diagnosis,for example:
> An OR gate A,with x,y as primary inputs and z as primary output,if gate is correct then z=x+y(z=x¡Åy).But if input is 0 and 1,output is 0,then we conclude A is faulty,this can be formalized as:
> AB(A)+(z=x+y) or ¦èAB(A)->(z=x¡Åy)
> where AB(A) denoting A is abnormal,we do not care what the output of the gate A will be if A is faulty.i.e.,z is unknown if A is not correct. to achieve this , I should attach to each gate an addtional primary input,when this primary input is 1,the gate act normally,otherwise output of the gate is unknown,can vis accept such specification and construct the BDD for this kind of gate? maybe someone can help me.thanks anyway
> Best regards,
> Yimin Zhang
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com
--
This archive was generated by hypermail 2.1.7 : Wed Nov 05 2003 - 01:31:50 MST