[SOLVED] Combining multipliers to make larger multiplier

[Wallboy]

I'm trying to create a 32x32 Unsigned Multiplier using smaller multipliers such as using 16x16 multipliers. My gut feeling is I need to use adders and shifters somewhere, but I just can't figure out how to do it. Any help would be appreciated, thanks.

 

[FvM]

Think about (a+b)(c+d) = ac + ad + bc + bd.

 

[Wallboy]

I seen your post in another thread and did it on paper, but it seems I would need a multiplier for each bit. I want something like two 16x16 multipliers to eventually get to a 32x32 result.

 

[FvM]

ac + ad + bc + bd suggests four 16x16 multipliers needed, isn't it?

 

[Wallboy]

Alright I think I'm starting to understand, I just need to figure out the shifts and partial add spots.

 

[tariq786]

FVM dude

what is (a+b) (c+d) here?

Are they upper and lower 16 bits of the multiplicand and the multiplier?

 

[FvM]

what is (a+b) (c+d) here?
Are they upper and lower 16 bits of the multiplicand and the multiplier?

Yes, exactly.

 

[Wallboy]

Alright, figured it out. Just needed four 16x16 multipliers and five 16x16 adders. No shifters were necessary.

 

[tariq786]

wallboy dude

please share the schematic with all of us here so that we also learn and reinforce concepts

thanks

 

[Wallboy]

Sure, I drew up a quick schematic that should work to build a 2Nx2N multiplier using four NxN multipliers.