extremely bad performance of event issuing by slider uicontrol in version R2020b

I have observed that in the new version of Matlab, the slider uicontrol does not issue events properly when used continuously. The value of the slider does not change as the user moves the slider.

I attach a function that illustrates this behaviour.

function sliderPerformance

function sliderPerformance
nfig = figure();
x    = linspace(0,2*pi,10000);

hplot = plot(x,cos(0*x),'LineWidth',2);
ylim([-2,2]);
h = uicontrol('style','slider','units','pixel','position',[20 20 300 20],'Parent',nfig);

% checks continuosly the update of the value in the slider
addlistener(h,'Value','PostSet',@(hObject, event) updatePlot(hObject, event,x,hplot));


function updatePlot(src,event,x,hplot)
persistent lastVisit
if isempty(lastVisit);
    lastVisit = clock();
    return
else
    thisVisit = clock();
end
fprintf('Time since last event: %5.2f s\n',etime(thisVisit,lastVisit));

hSlider = event.AffectedObject;
n    = hSlider.Value;
hplot.YData=cos(x*n);
ylim([-2,2]);
drawnow;
lastVisit = thisVisit;

When operated from any version previous to R2020b, the plot transitions smoothly as the user moves the slider, and (a sample of) the readout to the screen is:

Time since last event: 0.01 s

Time since last event: 0.11 s

Time since last event: 0.06 s

In R2020b, the plot transition is bumpy, and, accordingly the readout:

Time since last event: 0.08 s

Time since last event: 0.47 s

Time since last event: 1.11 s

Time since last event: 2.25 s

Time since last event: 4.51 s

Time since last event: 3.46 s

Time since last event: 4.65 s

Is something that can be done? Is this reportable as a bug?

Answer 1

[Cris LaPierre]

Your code runs smoothly for me in 20b and 21a prerelease.

Times are

• Time since last event: 0.04 s
• Time since last event: 0.04 s
• Time since last event: 0.04 s
• Time since last event: 0.03 s
• Time since last event: 0.04 s
• Time since last event: 0.04 s
• Time since last event: 0.04 s
• Time since last event: 0.05 s
• Time since last event: 0.05 s
• Time since last event: 0.04 s
• Time since last event: 0.03 s
• Time since last event: 0.03 s
• Time since last event: 0.04 s
• Time since last event: 0.03 s
• Time since last event: 0.04 s
• Time since last event: 0.04 s
• Time since last event: 0.04 s
• Time since last event: 0.04 s
• Time since last event: 0.04 s

[Arabarra]

Hi Cris,

very weird. I am using the same Mac (Mojave 10.14.6), for both my tests (smooth in 20a, terrible in 20b).

What is your system?

cheers,

D.

[Cris LaPierre]

I'm on the most recent version of Windows 10.

[Arabarra]

As an additional data point: for me, the error comes also in R2021a prerelease :-/:

Time since last event: 0.09 s

Time since last event: 1.65 s

Time since last event: 2.76 s

Time since last event: 1.47 s

Time since last event: 1.34 s

Time since last event: 0.77 s

Time since last event: 0.13 s

Time since last event: 0.27 s

Time since last event: 0.29 s

Time since last event: 0.22 s

Time since last event: 0.62 s

Time since last event: 1.31 s

Time since last event: 2.66 s

Time since last event: 1.90 s

[Cris LaPierre]

I tried it at home on a laptop from 2006 running Windows 7. It was a little slower, but still not like what you are reporting.

Time since last event: 0.09 s

Time since last event: 0.21 s

Time since last event: 0.11 s

Time since last event: 0.08 s

Time since last event: 0.08 s

Time since last event: 0.07 s

Time since last event: 0.13 s

Time since last event: 0.09 s

Time since last event: 0.11 s

Time since last event: 0.09 s

Time since last event: 0.17 s

Time since last event: 0.09 s

Time since last event: 0.73 s

Time since last event: 0.10 s

Time since last event: 0.11 s

Time since last event: 0.08 s

Time since last event: 0.12 s

Time since last event: 0.07 s

[Arabarra]

thanks Chris. Well, I guess I'll file it to Mathworks, there might be somehting in the Mac version that creates this problem.

[Arabarra]

Also, I just checked it on a Ubuntu 18.04 machine: same problem.

[Cris LaPierre]

If you are noticing a drastic difference in behavior from one version to another, I know they'd love to hear about it. Sory I couldn't be of more help. You can submit your feedback here.

[Arabarra]

I''ll do. Thanks Cris!