# Forum: FPGA, VHDL & Verilog FLoating point multiplier Logicore

 Author: Misbah Faiz (Company: fast) (mis6358) Posted on: 2014-05-21 20:17

Rate this post
 0 ▲ useful ▼ not useful
Does any one know the function of floating point multiplier in logicore?
its not giving me accurate answers i want to know how it works and how
it gives the answer and in what form

 Author: Lothar Miller (lkmiller) (Moderator) Posted on: 2014-05-21 21:20

Rate this post
 0 ▲ useful ▼ not useful
What manual do you use as reference? What information are you missing
there? What results do you expect and what results do you get?

 Author: Rolf S. (audiorolf) Posted on: 2014-05-21 22:32

Rate this post
 0 ▲ useful ▼ not useful
"Logicore" is Xilinx, right? You will have to provide code and an
example to make people be able to help you.

 Author: Misbah Faiz (Company: fast) (mis6358) Posted on: 2014-05-23 16:11

Rate this post
 0 ▲ useful ▼ not useful
I am using the manual which is provided on the xilinx site. A want a
answer of a simple multiplier like 1.2 x 1.2 should give me 2.4 in
binary at the output but it is giving me some dummy values.

 Author: Misbah Faiz (Company: fast) (mis6358) Posted on: 2014-05-23 16:12

Rate this post
 0 ▲ useful ▼ not useful
There is no code. I am using schematics. A simple floating point
multiplier in the schematics.

 Author: Klaus Falser (Guest) Posted on: 2014-05-23 16:18

Rate this post
 0 ▲ useful ▼ not useful
Misbah Faiz wrote:
> I am using the manual which is provided on the xilinx site. A want
> a
> answer of a simple multiplier like 1.2 x 1.2 should give me 2.4 in
> binary at the output but it is giving me some dummy values.

I'm not surprised.
1.2 x 1.2 gives 1.44 in my understanding.

 Author: AHED (Guest) Posted on: 2014-05-23 16:50

Rate this post
 0 ▲ useful ▼ not useful
I suppose

"1.2" means a signed fractional format with 1 integer bit and 2
fractional bits.
When multiplying two such numbers using a 3-bit signed integer
multiplier the result has to be interpreted as "2.4".

However, when dealing with floating point ( normalized mantissa, || <
1.0 )
the multiplication result is usually modified by dropping the unused
MSbits
so that "1.4" is left. Now, the new additional LSbits ( in this case 2 )
that cannot be represented in the destination format which is usually
the same as the input format "1.2" are dropped as well ( possibly after
rounding )

This might explain the confusion

• $formula (LaTeX syntax)$