decimal (floating point) multiply by integer number in vhdl

[TrickyDicky]

O(i) is a single bit. You still have not created an array of ufixed.
And what you are doing is way more complicated that it needs to be. You can do a rightshift without a clock.

 

[fanwel]

Yes, my input is an array of integer (with unspecified size) and I want to multiply the input with a fixed point. My previous post is just an example;
input, f= (12,3,10,7). Then, multiply the input with 0.25.

I never plays with decimal number(floating foint) before. I have read books/materials about this topic and I try to write a code for better understand. But, I become more confius now.
My idea is first to convert the integer array to ufixed since it cannot multiply directly wth 0.25. I'm sorry if I ask for a stupid question, I just want to learn. Thanks for reply

 

[FvM]

The below code does what you are asking for. But is doesn't make much sense in my opinion.

package my_data_types is
type exam is array (natural range <>) of integer;
end my_data_types;

Library IEEE;
use IEEE.std_logic_1164.all;
use work.my_data_types.all;
use IEEE.numeric_std.all;

entity fix is
port (A: in exam (1 to 8);
O: out exam (1 to 8));
end fix;

architecture fix of fix is
begin
process (a)
  begin
   for i in 1 to 8 loop 
     o(i) <= a(i)/4;
   end loop;  
  end process; 
end;

 

[fanwel]

Hi all;

Greatly thanks for your supports and I'll try the way you just told me.
Many thanks

 

[fanwel]

Hi all;

I have write a code using fixed_pkg. It compile successfully in quartus, but when I try to simulate in ModelSim an error occur.
---------------------------------------------------------------------------------------------------------------
library ieee;
library ieee_proposed;
use work.fixed_pkg.all;
use ieee.numeric_std.all;

entity fix is
port (clk: in bit;
nprev: in integer range -127 to 127;
ip1: out ufixed (8 downto -9));
end fix;

architecture fix of fix is
signal n1: ufixed (4 downto -4);
begin
process(clk)
begin
if (clk'event and clk='1') then
ip1 <= (to_ufixed(0.483,n1)) * (to_ufixed(11,n1));
end if;
end process;
end fix;
----------------------------------------------------------------------------------------------
This is the error in ModelSim:
# Compile of fix.vhd failed with 2 errors.
# Compile of fix_tb.vhd failed with 2 errors.
# Compile of fixed_float_types_c.vhdl was successful.
# Compile of fixed_pkg_c.vhdl failed with 1 errors.
# Compile of fixed_synth.vhdl failed with 2 errors.

I have add all the fixed_pkg in the project (in quartus). Need your helps..

---------- Post added at 08:49 ---------- Previous post was at 08:45 ----------

This the error when I expand it:
** Error: H:/altera/mul_try_tb/fix_tb.vhd(2): Library ieee_proposed not found.

 

[TrickyDicky]

in modelsim (and quartus because it ignored library names) you have to use floatfixlib:

library floatfixlib;
use floatfixlib.fixed_pkg.all;

Also, do NOT add any of the floatfixlib to the work library.

 

[fanwel]

I have add the two lines you give in quartus:

library ieee;
library ieee_proposed;
library floatfixlib;
use floatfixlib.fixed_pkg.all;
use ieee_proposed.fixed_pkg.all;
use ieee.numeric_std.all;

But the same error still occur..why, thanks for reply

 

[TrickyDicky]

also, remove the ieee_proposed library

what version of modelsim?

 

[fanwel]

Sorry, what do you mean by this word: Also, do NOT add any of the floatfixlib to the work library.

---------- Post added at 12:41 ---------- Previous post was at 12:31 ----------

I use ModelSim Altera Starter Edition 6.6

---------- Post added at 12:56 ---------- Previous post was at 12:41 ----------

After I remove the ieee_proposed library, this error occur in ModelSim:

# Compile of fix.vhd failed with 2 errors.
# Compile of fix_tb.vhd failed with 2 errors.
# Compile of fixed_float_types_c.vhdl was successful.
# Compile of fixed_pkg_c.vhdl failed with 1 errors.
# Compile of fixed_synth.vhdl failed with 2 errors.
** Error: H:/altera/mul_try_tb/fix.vhd(3): Library floatfixlib not found.
** Error: H:/altera/mul_try_tb/fix.vhd(4): (vcom-1136) Unknown identifier "floatfixlib".

 

[TrickyDicky]

ahh, I think the starter edition doesnt have it included, so forget my last post.

You will need to create a library called ieee_proposed in modelsim and add the fixed_pkg code to it.

 

[fanwel]

Do you mean add the fixed_pkg code under the ieee_proposed in modelsim (in one file)?
Thanks for reply