Graphs collapse with Matlab R2024a

Hi Professor,

I recently switched to a new Dell laptop with NVIDIA Gforce MX550 graphics card and Matlab R2024a installed.
It is weird that when runnning the following dynare file (see attached)
test.mod (13.1 KB)

the left two subplots of fig 1 do not display properly - they appear as 2 horizontal lines which wasn’t the case with other laptops.

I tried to configure my OpenGL setting, changing it from hardware to software, which didn’t seem to help…

Could you please provide some clues as to what has gone wrong - could is be resolved by configuring Matlab or related to updating Graphics card driver ( which I tried already but didn’t work either)…

Best,

This is not a plotting issue. You can inspect the plotted variables. They are indeed flat.

1 Like

Thanks for your prompt reply.
I think I know the reason now, as the Matlab installed in my new laptop automatically chose the vertical axis range to be [-1,1], which makes the responses flat lines…

Below is the figure produced by my other laptop where the vertical axis range was specified as [-0.06, 0.14]…

Now my question becomes how come there exists such discrepancy in the axis range automatically chose by Matlab?

Again, the contents of the two variables on the left on my machine are exactly zero. There must be something else going on.

Agree…
This time I modified the axis range but the horizontal lines still persist…

maybe has something to do with the Matlab version?
my other laptop has matlab 2022a installed and generated the desirable results.

The flat line appear with R2024 version.

That could be possible. You should investigate why the program returns different results. Did you use the same Matlab version on both machines?

1 Like

No, the new laptop that generated horizontal lines has Matlab R2024a and Dynare 6.1 installed, whereas the old one that generated desirable results has Matlab R2022a and Dynare 5.1 installed.

I also installed Matlab 2022a in the old one along with dynare 6.1 (as I cannot download the previous version) - the horizontal lines persist…so I suppose it might have something to do with the Dynare version?

Is there any way I can still download Dynare 5.1?

best

I think I found the issue. The differences in the steady state are within the standard tolerance of the solver. Try using

    steady(tolf=1e-10,tolx=1e-10); // ask Dynare to compute the ss for 500 values of chi
1 Like

It worked this time!!
Thanks for your diagnosis