Re: reduction operators (mantix 938)

Hi Shalom,

I dont remember what the very original query was - need to look at 
mantis, but i think it was with reference to support for unary reduction 
operators in analog (not bitwise as i have mentioned).

I was refering to AMS version 2.2, which already has unary reduction 
operators as part of the syntax and the language, but i dont think it 
has the entire set of operators as defined in 1364 or 1800. Only the &/| 
versions are defined in analog at present.

With the merger with 1364-2005, i would expect that there would be only 
one version and analog does not need a seperate set of these. We will 
just allow all the reduction operators defined in 1364 in the analog 
context also, from a syntax point of view. I think thats the intention, 
but Graham can probably comment on this.

Regards,
Sri

Bresticker, Shalom wrote:
> I'm not sure what version of the standard Sri was looking at.
> 
> The bit-wise binary operators are &, |, ^, ^~, ~^. These are binary
> operators, i.e., they take 2 operands.
> 
> There is 1 unary bit-wise operator: ~ (bit-wise negation, 1's
> complement).
> 
> The unary reduction operators are: ^, ^~, ~^, &, ~&, |, ~|. These take
> only one operator and the result is only one bit. Although some look the
> same as binary bit-wise operators, they are different ones. You know
> from context which is being used, by whether it is used as a binary or
> unary operator.
> 
> In Verilog, you can use all of these on integers, but not on real
> numbers.
> 
> Do you really need an analog version?
> 
> Shalom
>  
> 
> 
>>-----Original Message-----
>>From: owner-verilog-ams@eda.org [mailto:owner-verilog-
>>ams@eda.org] On Behalf Of edaorg@v-ms.com
>>Sent: Monday, January 30, 2006 11:16 AM
>>To: verilog-ams@eda.org
>>Subject: [Fwd] reduction operators (mantix 938)
>>
>>
>>From: Jonathan David <jb_david@yahoo.com>
>>
>>Hm.. maybe I should indicate a willingness to do more
>>than LURK on this list and help with the 2.3 draft,
>>even though I won't get Scintera to join Acellera.
>>
>>---Sri's last comments were:
>>I thought they were already present in the language as
>>bit wise operators - In chapter 4, Table 4-1 we have
>>operators such as ~& and ~| which is bitwise and/or
>>operators. Probably we dont have bitwise xnor.
>>
>>Is the same thing being referred here or something
>>else in digital LRM that we dont have?
>>
>>Need to have the full array of operators in digital.
>>
>>Some might have some restrictions - real?
>>-------------------------------------------
>>what I discovered (and mentioned to Martin a while
>>back) was that on the digital side for
>>reg [1:0] a;
>>
>>you get the following truth table for (&a)
>>a  &a  |a  ^a
>>00  0   0   0
>>01  0   1   1
>>10  0   1   1
>>11  1   1   0
>>
>>this is the essence of a reduction operator as opposed
>>to a bitwise operator..
>>the single (but multibit) argument is reduced to a
>>single bit..
>>
>>like their bitwise version, in the analog context the
>>domain should probably be limited to integers..
>>
>>Jonathan
>>
>>
>>
>>
>>
>>
>>----- End Included Message -----
> 
>