Hello Everyone, I am having trouble of late trying to print out noise summary reports following a Spectre noise analysis. The following problems seem to show up: 1) Noise contributors aren't printing correctly for the full hierarchy tree even though I am not selecting a particular hierarchy level when printing the noise summary. I am forced to select hierarchy level to something, e.g. "1" or "2" etc. else some noise contributors are printed but with zero contribution. This is obviously not right since the items printed would make a significant nonzero contribution, but is likely due to the next problem, which is: 2) Noise contributor numbers shown are unreasonably and impossibly large, e.g. 10^27 V^2. The percentage noise contributions (which by definition cannot exceed 100) are also astronomically large, e.g. 10^30 %. The total summarized noise is perfectly in order. So, it's not as if there is some modelling issue that might cause the noise to blow up in simulations. Given that I wasn't seeing any of the above problems until recently and there has been no change to either my Spectre binary release or to the models anytime in the last several weeks, I am guessing that there must be a rogue setting somewhere causing things to go wrong. Does anyone have an idea what the cause and resolution might be? The noise summary as it is now is totally unusable. I am using spectre release 19.10.063. Thanks, Vivek
↧