Blendics Improving Computer Scalability & Reliability › Forums › USE › MetaACElog.txt show errors… what do they mean?
September 27, 2016 at 1:50 pm #26957
The log file shows things like this:
Error in ==> find_metastability_time at 75
Error in ==> metaTimes at 4
Error in ==> MetaACE2 at 962
Error in ==> MetaACE2 at 2149
Error in ==> MetaACE2 at 1041
Where I can find some information about these errors?
September 27, 2016 at 1:51 pm #26959
Such errors are the result of some “odd” behavior of the circuit that MetaACE cannot figure out. While MetaACE can detect some common issues with the circuit and alert you with more useful errors, not every case can be done this way so these generic errors are printed. These do mean something to the developers so please report them so we can try to help.
For these, above, I suspect #2 and #3 are actually related to #1.
#1 is most likely a problem with the simulation not returning “good” data. I would check the following:
a) In the list file from the simulation, see if there are any HSPICE errors that indicate a problem with the netlist or simulation, itself. Fix them if you find them, of course.
b) If the list file is clean, then check the parameters/plots in MetaACE. Are you seeing the circuit behave as you expect? Do you see “full” simulations or do you see the circuit still metastable near the end of the simulation? Do you see both rising and falling transitions or do all traces resolve low or high? Are you looking at the correct node(s) in your circuit (i.e. ones that actually go metastable)?
It can be tricky to resolve all such cases, but allowing for a very wide (like 1 ns) setup window usually helps to ensure you are finding a good case that drives the circuit metastable. Also check the clock to be sure it’s not too narrow or really wide (or that it’s set properly for single or double edge as required by your circuit).
If you can get good-looking plots, this should resolve itself.
You must be logged in to reply to this topic.