>> There is an example of a 2x2 sequential multiplier. If we were to build a
>> 3x3 sequential multiplier, all we'd have to do is add one of each
>> component
>> right? For instance, add a D-flip flop for each operand and for the
>> sum-storage, add an AND gate, and add a FA.
>
> Right. How much longer would it take to finish computing?

Actually, I think we have to add two D-flip-flops to the second row, since
we will be shifting that operand twice (in the 2x2 case, we only needed one
extra D-flip-flop since we are shifting that operand once). We would also
need two extra AND gates, two extra FA, and two extra D-flip-flops to store
the sum-terms (at the bottom).

To answer your question, would it take one extra cycle to compute the
results (assuming the operands are already in the registers)?

Thanks,
RC

----- Original Message -----
From: "Robert Dick" <dickrp@northwestern.edu>
To: <ADLD@listserv.it.northwestern.edu>
Sent: Wednesday, December 10, 2008 3:39 PM
Subject: Re: Fw: questions

>> CLA
>> In the slides for Carry Lookahead Adder, the equations are
>>
>> cin1 = Go + Po cout0
>> cin2 = G1 + P1 cout1 = G1 + P1G0 + P1P0 cout0
>> etc
>>
>> Shouldn't it be this:
>> cin1 = Go + Po cin0
>> cin2 = G1 + P1 cin1 = G1 + P1G0 + P1P0 cin0
>
> Yes, it should. I fixed that slide (cout->cin) and reposted to the
> website.
>
>> Multipliers
>> In the Cascaded Carry Lookahead diagram, are the four 4-bit adders CLAs
>> as
>> well? If not, they can be right?
>
> They can be CLAs or even CSAs.
>
>> There is an example of a 2x2 sequential multiplier. If we were to build a
>> 3x3 sequential multiplier, all we'd have to do is add one of each
>> component
>> right? For instance, add a D-flip flop for each operand and for the
>> sum-storage, add an AND gate, and add a FA.
>
> Right. How much longer would it take to finish computing?
>
> -Robert Dick-
>
Received on Wed Dec 10 15:59:24 2008