I can provide an example of this bug. Here is how to reproduce the issue:
1. Download the public 3.0.0 release of the U.S. Energy Policy Simulator at
https://github.com/Energy-Innovation/ep ... /3.0.0.zip
2. Open the EPS.vpmx file (NOT the EPS.mdl) file in Vensim DSS.
3. Click Sim Control > Changes tab > Load changes from... button and select the file "Scenario_NetZero.cin" that is included in the model folder.
4. Click the "Simulate" button that is inside the "Sim Control" pop-up box (not the "Simulate" button on the main toolbar)
5. Vensim reports that it's reading from "Scenario_NetZero.cin," runs the model, and produces results. Hit the "Table" button. Unless you have accidentally clicked on some other variable, the variable "Output Total CO2e Emissions" should be selected, and it should give you a table of numerical results. Notice the result in 2050 is -1.74853
6. If you want to verify that the "Scenario_NetZero.cin" file is still selected, you may open "Sim Control" again and look at the "Changes" tab to verify the changes file is still selected. Hit "Set" to close the "Sim Control" pane. This step is optional, as it is only to prove that we're still going to run with the same exact .cin file.
7. Click the "Simulate" button on the main toolbar.
8. Vensim reports that it's reading from "Scenario_NetZero.cin," runs the model, and produces results. Hit the "Table" button. Unless you have accidentally clicked on some other variable, the variable "Output Total CO2e Emissions" should be selected, and it should give you a table of numerical results. Notice the result in 2050 is -1.57735 (not -1.74853)
In short, the model produces different numerical results depending on whether you run the model by using the "Simulate" button in the main toolbar, or the "Simulate" button in the "Sim Control" pane. While the result is only slightly different in the example above, we've encountered cases where the results are completely, dramatically different depending on which "Simulate" button you use.
You can actually toggle back and forth between the two sets of results by running the model using one Simulate button or the other, with the same .cin file remaining selected.
Of course, the model results ought not to be different depending on which Simulate button you click.
This bug only happens with the .vpmx file. If you perform the same test above but using the .mdl file, you will get a result of -1.57735 when using either Simulate button.