"Real" Simulations?

[Jay Davis]

I'm currently using Xilinx Webpack and modelsim to develop a project
(yes, it's a class assignment). The problem is, I'm not allowed to take
the FPGA board out of the lab, and I'd like to work on some of the
project at home. I was wondering if there exists any sort of software
that would be able to drive the parallel port of a PC with the outputs
of a VHDL simulation? In other words, would it be possible to use a
simulation to drive some logic values onto the parallel port which I
could then pretend were the outputs of my FPGA chip (and then use those
outputs to test my VHDL code with the rest of my circuit)? I realize
that I would have some limitations on speed, but it ought to be
theoretically possible (or maybe not, that's why I'm asking!).

Thanks for any pointers!

Jay

 

[Mike Treseler]

Jay Davis wrote:

I'm currently using Xilinx Webpack and modelsim to develop a project
(yes, it's a class assignment). The problem is, I'm not allowed to take
the FPGA board out of the lab, and I'd like to work on some of the
project at home. I was wondering if there exists any sort of software
that would be able to drive the parallel port of a PC with the outputs
of a VHDL simulation?

Consider watching those outputs as modelsim waves
and then close the loop and have your testbench
verify the expected outputs.

-- Mike Treseler

 

[Jay Davis]

Mike Treseler wrote:

Consider watching those outputs as modelsim waves
and then close the loop and have your testbench
verify the expected outputs.

I'm sorry, I don't understand what you mean by "close the loop". I
don't want to write code to drive a parallel port, I want a program that
can drive a parallel port with the output waveform of a modelsim run.
In this way, I would not need physical access to a FPGA chip to test my
design with my other hardware - I could use the PC instead.

--
Jay

 

[Jonathan Bromley]

"Jay Davis" <no@spam.com> wrote in message
news:vXivb.429855$be.60632@news.easynews.com...

Consider watching those outputs as modelsim waves
and then close the loop and have your testbench
verify the expected outputs.

I'm sorry, I don't understand what you mean by "close the loop". I
don't want to write code to drive a parallel port, I want a program that
can drive a parallel port with the output waveform of a modelsim run.
In this way, I would not need physical access to a FPGA chip to test my
design with my other hardware - I could use the PC instead.

Mike is pushing you towards a professional's view of
simulation:

1) Use the simulator as a debug tool, to help you get the
design basically working. At this stage you are likely
to make heavy use of the waveform view, code tracing and
breakpoints, and other debug features of the simulator
that help you probe around inside your code.
2) Once your design is showing real signs of life, you can
start to use the simulator as a *verification* tool. To
do this properly, you need to get the testbench to predict
what the outputs should be for any given stimulus.
Then the testbench itself can compare the expected output
with the simulated design's output, and report any
discrepancies - probably in the form of text messages
from assert statements.

Whether you are ready for this depends on how confident you
are in writing VHDL for testbenches - a very different art
from writing designs.

To go back to your original question: No, I don't know of
any simulation tools that allow you to use real PC peripherals
to mimic I/O in your simulated system. In any case, it would
be desperately limiting to have only the PC's I/O facilities -
the great thing about simulation is that you are limited only
by your imagination. However, a common approach to this kind
of problem is to mock-up your external system (a few buttons
and lights, in your case?) using Tcl/Tk, Visual Basic or some
similar GUI-creation system. Then you need to find some way
to get this system mock-up to talk to your VHDL simulator.
If you have the full (SE) version of ModelSim, that's rather
easy because you can write Tcl/Tk GUIs within ModelSim itself;
otherwise, it's a bit more tricky because you need to have
two independent programs talking to one another. Either
way, it's sure to be harder work than writing a classic
self-checking VHDL testbench as Mike suggested.
--
Jonathan Bromley, Consultant

DOULOS - Developing Design Know-how
VHDL * Verilog * SystemC * Perl * Tcl/Tk * Verification * Project Services

Doulos Ltd. Church Hatch, 22 Market Place, Ringwood, Hampshire, BH24 1AW, UK
Tel: +44 (0)1425 471223 mail: jonathan.bromley@doulos.com
Fax: +44 (0)1425 471573 Web: http://www.doulos.com

The contents of this message may contain personal views which
are not the views of Doulos Ltd., unless specifically stated.

 

[Charles M. Elias]

Jay Davis <no@spam.com> wrote in message news:<77hvb.423205$Te.3094381@news.easynews.com>...

I'm currently using Xilinx Webpack and modelsim to develop a project
(yes, it's a class assignment). The problem is, I'm not allowed to take
the FPGA board out of the lab, and I'd like to work on some of the
project at home. I was wondering if there exists any sort of software
that would be able to drive the parallel port of a PC with the outputs
of a VHDL simulation? In other words, would it be possible to use a
simulation to drive some logic values onto the parallel port which I
could then pretend were the outputs of my FPGA chip (and then use those
outputs to test my VHDL code with the rest of my circuit)? I realize
that I would have some limitations on speed, but it ought to be
theoretically possible (or maybe not, that's why I'm asking!).

Thanks for any pointers!

Jay
Jay,

Make a VHDL model of the "rest of your circuit"--I'll call it
device_b. Make a component of your FPGA device (device_a) and
instantiate it in a testbench for device_b. Use the outputs of
device_a (device_a is driven by the inputs you used to test it) as
inputs to device_b. You "close the loop" by monitoring device_b's
outputs and checking them for correctness either by inspection or by
writing VHDL code to verify the outputs.

The answer to your question about using the parallel port is: It
depends. It is theoretically possible if you are not asking for too
many signal outputs, but I don't know of any software that would do
the job for you. Since your time is probably limited, the testbench
solution is more practical, I think.

Best regards,

Charles

 

[Jay Davis]

Charles M. Elias wrote:

Make a VHDL model of the "rest of your circuit"--I'll call it
device_b. Make a component of your FPGA device (device_a) and
instantiate it in a testbench for device_b. Use the outputs of
device_a (device_a is driven by the inputs you used to test it) as
inputs to device_b. You "close the loop" by monitoring device_b's
outputs and checking them for correctness either by inspection or by
writing VHDL code to verify the outputs.

Ah, I understand now. This would work if the "rest of my circuit" was
simple (just a few buttons and LEDs, as another poseter suggested), but
what I'm doing is interfacing a (256x128) LCD, and modeling an LCD in
VHDL isn't going to be practical. I have my code already generating the
right sequence of commands to do what I want (judging by the testbench
waveform output), but it's difficult to be sure that I have all the
timing issues correctly sorted out when I only have physical access to
the FPGA for three hours a week.

The answer to your question about using the parallel port is: It
depends. It is theoretically possible if you are not asking for too
many signal outputs, but I don't know of any software that would do
the job for you. Since your time is probably limited, the testbench
solution is more practical, I think.

All right, that's what I figured. Sounds like I'll just do the best I
can and petition for more lab time!

Thanks,
Jay

 

[Brian Drummond]

On Fri, 21 Nov 2003 17:57:52 GMT, Jay Davis <no@spam.com> wrote:

Charles M. Elias wrote:

Make a VHDL model of the "rest of your circuit"--I'll call it
device_b. Make a component of your FPGA device (device_a) and
instantiate it in a testbench for device_b. Use the outputs of
device_a (device_a is driven by the inputs you used to test it) as
inputs to device_b. You "close the loop" by monitoring device_b's
outputs and checking them for correctness either by inspection or by
writing VHDL code to verify the outputs.

Ah, I understand now. This would work if the "rest of my circuit" was
simple (just a few buttons and LEDs, as another poseter suggested), but
what I'm doing is interfacing a (256x128) LCD, and modeling an LCD in
VHDL isn't going to be practical.

It might be. To model it exhaustively would be difficult.

But to accept all the data fed to it (only 32kbits/frame?), or even only
the top few rows to save simulation time, and verify (or log to disk) a
small subset of that data (even an 8*8 cell in one corner) would be much
easier, and could catch a lot of potential problems.

- Brian

 

[Thomas Stanka]

Jay Davis <no@spam.com> wrote:

I'm currently using Xilinx Webpack and modelsim to develop a project

project at home. I was wondering if there exists any sort of software
that would be able to drive the parallel port of a PC with the outputs
of a VHDL simulation? In other words, would it be possible to use a
simulation to drive some logic values onto the parallel port which I
could then pretend were the outputs of my FPGA chip (and then use those
outputs to test my VHDL code with the rest of my circuit)? I realize
that I would have some limitations on speed, but it ought to be
theoretically possible (or maybe not, that's why I'm asking!).

Modelsim is able to to drive your parallelport via the FLI (foreign
language interface). With the fli you can insert C-code in your
simulation and drive the parport. It is not the easiest way and may
took some time until you manage to get signals via the parport if your
not familiar using the parport from C.
I think our student apprentice took 3-4 weeks until he was able to use
the parport with modelsim in both directions.

bye Thomas

 

[Tim Hubberstey]

Thomas Stanka wrote:

Jay Davis <no@spam.com> wrote:
I'm currently using Xilinx Webpack and modelsim to develop a project

project at home. I was wondering if there exists any sort of software
that would be able to drive the parallel port of a PC with the outputs
of a VHDL simulation? In other words, would it be possible to use a
simulation to drive some logic values onto the parallel port which I
could then pretend were the outputs of my FPGA chip (and then use those
outputs to test my VHDL code with the rest of my circuit)? I realize
that I would have some limitations on speed, but it ought to be
theoretically possible (or maybe not, that's why I'm asking!).

Modelsim is able to to drive your parallelport via the FLI (foreign
language interface). With the fli you can insert C-code in your
simulation and drive the parport. It is not the easiest way and may
took some time until you manage to get signals via the parport if your
not familiar using the parport from C.
I think our student apprentice took 3-4 weeks until he was able to use
the parport with modelsim in both directions.

Unfortunately, the FLI is only available on the SE (super expensive :^)
) version of ModelSim.

A possible brute-force method is to use ModelSim's Tcl scripting to
execute an external (DOS) program each time you want to update the state
of your parallel port. This is likely to be obscenely slow but it can be
implemented with all versions of ModelSim. The Tcl commands you will
need to read up on are "when" and "exec".
--
Tim Hubberstey, P.Eng. . . . . . Hardware/Software Consulting Engineer
Marmot Engineering . . . . . . . VHDL, ASICs, FPGAs, embedded systems
Vancouver, BC, Canada . . . . . . . . . . . http://www.marmot-eng.com