configurations and generics

[Andy Luotto]

I am designing a verification environment with parameters, e.g.

entity cordic_tb is
generic(
pipe : integer;
datawidth : integer;
anglewidth : integer
);
end cordic_tb;
architecture beh of cordic_tb is ...

and i would like to design configurations of the same test bench with
different generics, something like

configuration c_cordic_v7_nom of cordic_tb is
for beh
-- generic map(
-- pipe => 5
-- datawidth =>11
-- anglewidth =>16
-- );

for i_cordic: cordic
use configuration bblocks_lib.cordic_v7_configuration;
end for;
.....

can i place the generic map (clike that ommented out) to control the
generics of the top (and not just the sub modules generics)?

i know i can pass those parameters as simulator arguments (e.g. ncelab
anglewidth=16 ...), still i'd like to handle this in the HDL to make
simualtor independent

thanks

 

[Andy]

On Sep 10, 9:45 am, Andy Luotto <andyluo...@excite.com> wrote:

I am designing a verification environment with parameters, e.g.

entity cordic_tb is
  generic(
    pipe  : integer;
    datawidth  : integer;
    anglewidth : integer
  );
end cordic_tb;
architecture beh of cordic_tb is ...

and i would like to design configurations of the same test bench with
different generics, something like

configuration c_cordic_v7_nom of cordic_tb is
  for beh
--  generic map(
--    pipe  => 5
--    datawidth  =>11
--    anglewidth =>16
--  );

    for i_cordic: cordic
      use configuration bblocks_lib.cordic_v7_configuration;
    end for;
....

can i place the generic map (clike that ommented out) to control the
generics of the top (and not just the sub modules generics)?

i know i can pass those parameters as simulator arguments (e.g. ncelab
anglewidth=16 ...), still i'd like to handle this in the HDL to make
simualtor independent

thanks

I think you may have to have a top level wrapper over the tb to do
that. But I don't use configurations much anymore, preferring to use
generics and entity instantiations. If you really want tool
independence, you can create a top level wrapper for each would-be
configuration that sets the generics for you.

Andy

 

[Mark McDougall]

Andy Luotto wrote:

and i would like to design configurations of the same test bench with
different generics, something like

I know this isn't any help, but good luck with that! I struggled with
getting configurations to do anything *useful* (in a concise manner) for
me and gave up after a couple of days...

Regards,

--
Mark McDougall, Engineer
Virtual Logic Pty Ltd, <http://www.vl.com.au>
21-25 King St, Rockdale, 2216
Ph: +612-9599-3255 Fax: +612-9599-3266

 

[Andy Luotto]

I know this isn't any help, but good luck with that! I struggled with
getting configurations to do anything *useful* (in a concise manner) for
me and gave up after a couple of days...

Regards,

on my side i can say i was *forced* by my customer to use vhdl, since
i largely prefers (system) verilog: half the time to do the same
stuff ... i tried to use vhdl in a concise manner which is competitive
with verlog with no success ;-)

 

[Andy]

On Sep 11, 2:17 am, Andy Luotto <andyluo...@excite.com> wrote:

I know this isn't any help, but good luck with that! I struggled with
getting configurations to do anything *useful* (in a concise manner) for
me and gave up after a couple of days...

Regards,

on my side i can say i was *forced* by my customer to use vhdl, since
i largely prefers (system) verilog: half the time to do the same
stuff ... i tried to use vhdl in a concise manner which is competitive
with verlog with no success ;-)

Using VHDL generics and entity instantiations can obviate
configurations and component declarations in all but the most
demanding, complex applications. These two items seem to be the most
egregious offenders when it comes to getting real work done in vhdl
code.

Andy