asynchronous reset, simulator doesn't support

[Clunixchit]

Hello there,

Here is a challenge for all of us.

I have a simulator (proof from alliance) that doesn't support
asynchronous reset. Alliance can work with using only one clock (and
not any synchronous set or reset).

Nevertheless I want to simulation asynchronous reset dff from its
behavioral description.

What would you recommend me to change in my .vhd file in terms of
asynchronous reset ?


-- D flip flop with asynchronous reset
library IEEE;
use IEEE.std_logic_1164.all;

entity dff_asynchrone is
port (
data, re_set, cp : IN std_logic ;
q : OUT std_logic
);
end dff_asynchrone;

architecture behave of dff_asynchrone is
begin
process(cp,re_set)
begin
if re_set = '0' then
q <= '0';
elsif (cp='1' and cp'event) then
q <= data;
end if;
end process;
end behave;

 

[KJ]

Hello there,

Here is a challenge for all of us.

I have a simulator (proof from alliance) that doesn't support
asynchronous reset. Alliance can work with using only one clock (and
not any synchronous set or reset).

Nevertheless I want to simulation asynchronous reset dff from its
behavioral description.

What would you recommend me to change in my .vhd file in terms of
asynchronous reset ?

I'd recommend keeping the .vhd file and changing your simulator. For a
simulator to 'not support asynchronous resets' really means that it doesn't
support 'if/elsif/end if' VHDL statements. Why waste time with such a tool?

KJ

 

[David Binnie]

It may be the target device, some devices do not support asynchronous
reset..

 

[Clunixchit]

I'd recommend keeping the .vhd file and changing your simulator. For a
simulator to 'not support asynchronous resets' really means that it doesn't
support 'if/elsif/end if' VHDL statements. Why waste time with such a tool?

KJ

I've modelsim and ghdl at my disposal as alternative simulation tools.
However it does support 'if/elsif/end if' VHDL statements, but not
asynchronous reset. The counter (see code below) does simulate
successfully.

My question wasn't about time, but about possible ways of coding
asynchronous reset.


entity count1 is
port ( clk, clear : in bit;
qc : out integer range 0 to 255
);
end count1;

architecture behavioral of count1 is
begin
process (clk)
variable cnt : integer range 0 to 255;
begin
if (clk'EVENT and clk= '1') then
if clear = '0' then
cnt := 0;
else
cnt := cnt +1;
end if;
end if;
qc <= cnt;
end process;
end behavioral;

 

[KJ]

I've modelsim and ghdl at my disposal as alternative simulation tools.
However it does support 'if/elsif/end if' VHDL statements, but not
asynchronous reset.

Both of those tools do support 'asynchronous reset' (aka if/elsif/end if)

The counter (see code below) does simulate
successfully.

My question wasn't about time, but about possible ways of coding
asynchronous reset.

Your original post (entity dff_asynchrone) is one correct way of coding an
asynchronous reset. Why do you think it.they do not simulate correctly?

KJ

 

[Clunixchit]

Both of those tools do support 'asynchronous reset' (aka if/elsif/end if)

I was referring to alliance here. it doesn't support 'asynchronous
reset'. but it does support the counter (see code in previous emails)

Your original post (entity dff_asynchrone) is one correct way of coding an
asynchronous reset. Why do you think it.they do not simulate correctly?

Again I meant alliance doesn't simulate correctly. Modelsim and ghdl
works.
It fails with:
Compiling 'dff_asynchrone' ...
VHDL : Error - bad usage of the 'stable' attribut
*** Compilation aborted...
make: *** [PROOF] Error 255

By referring Modelsim and ghdl above, I meant that I can use either
Modelsim or ghdl for my simulations. But it's curiosity on for
alliance that drove me to start this topic here.

regards,
Chitlesh

 

[KJ]

Both of those tools do support 'asynchronous reset' (aka if/elsif/end if)

I was referring to alliance here. it doesn't support 'asynchronous
reset'. but it does support the counter (see code in previous emails)

Your original post (entity dff_asynchrone) is one correct way of coding
an
asynchronous reset. Why do you think it.they do not simulate correctly?

Again I meant alliance doesn't simulate correctly. Modelsim and ghdl
works.
It fails with:
Compiling 'dff_asynchrone' ...
VHDL : Error - bad usage of the 'stable' attribut
*** Compilation aborted...
make: *** [PROOF] Error 255

By referring Modelsim and ghdl above, I meant that I can use either
Modelsim or ghdl for my simulations. But it's curiosity on for
alliance that drove me to start this topic here.

regards,
Chitlesh

Then re-read my original post about not using such a simulator. Next open a
service request against Alliance and tell them you won't be using their
lame tool.

KJ

 

[HT-Lab]

Both of those tools do support 'asynchronous reset' (aka if/elsif/end if)

I was referring to alliance here. it doesn't support 'asynchronous
reset'. but it does support the counter (see code in previous emails)

Your original post (entity dff_asynchrone) is one correct way of coding
an
asynchronous reset. Why do you think it.they do not simulate correctly?

Again I meant alliance doesn't simulate correctly. Modelsim and ghdl
works.
It fails with:
Compiling 'dff_asynchrone' ...
VHDL : Error - bad usage of the 'stable' attribut
*** Compilation aborted...
make: *** [PROOF] Error 255

By referring Modelsim and ghdl above, I meant that I can use either
Modelsim or ghdl for my simulations. But it's curiosity on for
alliance that drove me to start this topic here.

I would agree with KJ, this looks like a serious bug in the Alliance VHDL
parser, it complains about the 'stable attribute but you are using 'event?

I actually question this bug since if the parser can't handle 'event then
probably most designs will fail, do you get the same error with rising_edge?

Hans
www.ht-lab.com

 

[Andy]

On Sep 3, 4:05 pm, "KJ" wrote:
I was referring to alliance here. it doesn't support 'asynchronous
reset'. but it does support the counter (see code in previous emails)

Again I meant alliance doesn't simulate correctly. Modelsim and ghdl
works.
It fails with:
Compiling 'dff_asynchrone' ...
VHDL : Error - bad usage of the 'stable' attribut
*** Compilation aborted...
make: *** [PROOF] Error 255

By referring Modelsim and ghdl above, I meant that I can use either
Modelsim or ghdl for my simulations. But it's curiosity on for
alliance that drove me to start this topic here.

I would agree with KJ, this looks like a serious bug in the Alliance VHDL
parser, it complains about the 'stable attribute but you are using 'event?

I actually question this bug since if the parser can't handle 'event then
probably most designs will fail, do you get the same error with rising_edge?

Hanswww.ht-lab.com

regards,
Chitlesh

Perhaps this tool is a cycle (clock) based simulator? And maybe the
parser only accepts rising_edge() or falling_edge() calls for clock
edge "detection". Change your "(cp = '1' and cp'event)" to
"rising_edge(cp)". IINM, 'stable can be a more general case of the
opposite of 'event, and could be being used for it.

If this is really a true cycle based simulator, you're pretty much out
of luck trying to code an asynchronous reset, the simulator can
probably only handle things happening at the next clock edge. You
could convert to synchronous reset:

process(cp)
begin
if rising_edge(cp) then
if re_set = '0' then
q <= '0';
else
q <= data;
end if;
end if;
end process;

Andy

 

[Clunixchit]

I actually question this bug since if the parser can't handle 'event then
probably most designs will fail, do you get the same error with rising_edge?

Yes I still have the error.

If this is really a true cycle based simulator, you're pretty much out
of luck trying to code an asynchronous reset, the simulator can
probably only handle things happening at the next clock edge. You
could convert to synchronous reset:

So be it. As for a synchronous reset, the simulator does the job.

Thanks for your help.

Chitlesh