to david bishop

hi i have to implement a hopfield network with floating point numbers
using ise foundation tool 6.3i but i haven't any package of floating
point, it has any solution for this problem?thanks

dilou

dilou wrote:
> hi i have to implement a hopfield network with floating point numbers
> using ise foundation tool 6.3i but i haven't any package of floating
> point, it has any solution for this problem?thanks

http://www.eda.org/vhdl-200x/vhdl-20...ges/files.html
but there were incompatibilities with ISE.
Try ISE 8.1 or Quartus.

-- Mike Treseler

Mike Treseler

Mike Treseler wrote:
> dilou wrote:
>> hi i have to implement a hopfield network with floating point numbers
>> using ise foundation tool 6.3i but i haven't any package of floating
>> point, it has any solution for this problem?thanks
>
> http://www.eda.org/vhdl-200x/vhdl-20...ges/files.html
> but there were incompatibilities with ISE.
> Try ISE 8.1 or Quartus.

You said it, I didn't.

Hint on any Xilinx XST, comment out the divide routines.
You can substitute the NR_ (Newton Raphson) ones from the float_alg_pkg
for the divide routine if you really need it.

David Bishop

David Bishop

hi i don't understand this answer.....in my university, it has only the
ise foundation 6.3i, it is possible to order from xilinx through my
university a floating point package with a recent version of ise
foundation?and will my design work with this specifications?

dilou

dilou

dilou wrote:
> hi i don't understand this answer.....in my university, it has only the
> ise foundation 6.3i, it is possible to order from xilinx through my
> university a floating point package with a recent version of ise
> foundation?and will my design work with this specifications?

The floating point type looks like this:

type float is array (INTEGER range <>) of STD_LOGIC;
subtype float32 is is float (8 downto -23);

ISE foundation 6.3 can now work with a negative index
ISE 8.1 can.

David Bishop