VHDL200x- Fixed Point Problem in Quartus 5.0

[Divyang M]

Hi,

When I use the VHDL200x Fixed Point package in Quartus 4.2 SP1, it
works fine and I get correct results for my design (though there are 3
warnings and some Infos resulting from the package), but in Quartus 5.0
some conditions from the fixed point package evaluate to a constant and
the same design gives wrong results.

How can I go about using the fixed_point package successfully in
Quartus 5.0?

The details of the warnings and infos are below:

Fixed Point Package : version 1.3 2005-04-15

The warnings I get in Quartus 5.0 are :
Warning : (10037) Verilog HDL or VHDL warning at fixed_pkg_c.vhd (1566)
: condition expression evaluates to a constant

Same problem that occurs on line 1566 occurs at the following lines
also:
4042, 1754, 1666, 1525, 4042, 1754, 1173, 1177, 1713

In addition, I also get the following info:
Info: VHDL Case Statement information at fixed_pkg_c.vhd(3927): OTHERS
choice is never selected.

In Quartus version 4.2 SP1 I get the following warnings:
Warning: VHDL Subtype or Type Declaration warning at
fixed_pkg_c.vhd(1130): subtype or type has null range.

Same warnings on lines 1131, 1132

And same Info as Quartus 5.0.

Thanks,

Divyang M.

 

[Divyang M]

NOTE : The above is all actual circuit running on the FPGA. I used
ModelSim for simulation and the simulation is fine.

 

[Mike Treseler]

Divyang M wrote:

How can I go about using the fixed_point package successfully in
Quartus 5.0?

Report the problem to Altera
and send them your files.

Run your testbench on the .vho file
for both Quartus versions
to rule out a synthesis error.

-- Mike Treseler

 

[Divyang M]

Report the problem to Altera
and send them your files.

Will do so.

Run your testbench on the .vho file
for both Quartus versions
to rule out a synthesis error.

I am not sure if I get this. As I understand, .vho files are only for
simulation. I am using ModelSim (stand-alone version) for simulation.
Does the above mean I should run my simulation in both the version of
Quartus and check the results (How does this help me rule out synthesis
errors?) ? OR Is it enough to run simulation in the stand-alone
version?

Thanks,

Divyang M.

 

[Mike Treseler]

Divyang M wrote:

I am not sure if I get this. As I understand, .vho files are only for
simulation. I am using ModelSim (stand-alone version) for simulation.

With modelsim, you can run your testbench either
on your source code (.vhd) written by you,
or the synthesis netlist (.vho)written by quartus.

If your source code passes, but the netlist
doesn't, synthesis is suspect.

Maybe the 4.2 netlist passes and 5.0 netlist does not.

-- Mike Treseler

 

[David Bishop]

Divyang M wrote:

Hi,

When I use the VHDL200x Fixed Point package in Quartus 4.2 SP1, it
works fine and I get correct results for my design (though there are 3
warnings and some Infos resulting from the package), but in Quartus 5.0
some conditions from the fixed point package evaluate to a constant and
the same design gives wrong results.

How can I go about using the fixed_point package successfully in
Quartus 5.0?

e-mail me your code and I'll see what I can do. I did my testing on
this code with Quartus 4.2, and I have not tested it with Quartus 5 yet.

David Bishop dbishop@vhdl.org

 

[Divyang M]

Hi David,

Would it be more convenient for you if I send only the files (2 of
them) that use the fixed_package

OR

the whole project (about 6 files) with the pin assignment and all as
well as ModelSim simulation testbench and testvectors?

Thanks,

Divyang M.