Simulation and Verification

[jkvam]

For a few designs I've ISim will sometimes output weird looking traces (see picture). I'm wondering if anyone else has seen this behavior. I've seen this behavior on 13.3 (earliest I tested), 13.4, and now 14.1 on win7x64 using both 64 and 32-bit ISE tools. I have another machine with the same OS and I don't see this behavior.

I filed a webcase but the resolution was unhelpful, so I thought I would ping the community.

[hirschdaumen]

I know that this happening when the "0 ps" point of time is visible in the wave window. I think ISIM gets confused with the uninitialized signals and displays some kind of overlay of all states that can be seen in the actual view, or just 'X' or other junk. Scroll a bit to the right so that the simulation begin is not in the view anymore and the signal should be displayed ok.

[scampbell]

I think that in some cases it has been seen that the waveform viewer with ISim does produce waveforms like you show. The waveform viewer has been re-architected in the next generation software, and is a common waveform viewer with Chipscope to aid in simulating both functional and post-implementation results.

There are still ways to obtain the correct value at a certain point of time in the simulation. These are:

1.) looking at the value in the "values" column of the waveform

2.) Using the dump command to dump all values at the current simulation time

3.) Put value checking assertions in your testbench to print, or test, the values at certain simulation points

Finally, if everything else fails, ISim can output a standard Value Change Dump (VCD) for use by third party simulation viewers, some free such as GTKWave:

http://gtkwave.sourceforge.net/

OK, I hope this helps ...

[mlaffin]

I have had similar glitches in windows 7, ISE13.4 in 32 bit win7 and ISE14.2 in 64 bit win7. In the worst cases they completely crash ISIM in a flickering mess of error mesages. They are also random, ocasionally not occuring at all. After much frustration I just turned to, the unsupported, ubuntu operating system.  (Actually an old version, 10.04, 32 bit, pae, hyperthreading OFF.) I am not experienceing the glich for the exact same project files. You may want to consider an alternative operating system. (If you go with an unsupported OS the cable drivers may need some work arounds, see #http://www.george-smart.co.uk/wiki/Xilinx_JTAG_Linux, but I can program the ML605 for my project.)

[muzaffer]

I have seen cases where isim would show a signal with overlapping red and green or red ony lines and would read 1 for overlapping regions, X for all red when there was no reason for X. A VCD of the same waveform would show the waveforms as simulation as intended. So there are quite a few problems with ISIM graphical output but that's not the biggest problem. If I could get it to restart properly I can live with some graphical glitches (at least for a while longer ;-)