I'm (slowly) building a WordPress site with some didactic simulation made with Mathematica.

Users can run the simulations (on-line or off-line) with the free Wolfram CDF Player.

One of the last simulation I've made was about waves propagation and can be viewed here.

The notebook file (.nb) (rather long) can be downloaded here.

The cdf file uploaded in the site was created with MMA9 and can be downloaded here.

For those not having the CDF player installed on their PC there is a short youtube preview simulation here.

And here follows a static image of the simulation:enter image description here

After pressing the play button users can see the dynamic progress of the wave (there are three 2D plots showing that).

Furthermore, by checking the "Show 3D graph", the fourth 3D plot is displayed, showing the wave function. But enabling that option slows down the refresh rate of the whole dynamic output.

The oddities are the following:

  1. If I run the code with MMA9.0.1 or my CDF Player (as a standalone program) the simulation is acceptable, also with the 3D plot enabled (in this case the plots are refreshed about every 0.3/0.4 secs). Not very smooth but smooth enough to give the idea.
  2. If I run the code with MMA10.0.1 or through the browser the simulation with the 3D plot enabled is twice as slow (with the plots refreshed about every 0.7/0.8 secs): too much jerky!.

I'm well aware that the code I've come up with might not be perfectly optimized (that would be another kind of question) but here the problem that is puzzling me is that the same simulation (created with the same code) performs very differently in different running conditions: in Mathematica 10.0.1 and in my site (through the browser) is too slow, while in Mathematica 9.0.1 or in the CDF Player (standalone) is almost ok.

Another quandary is that, when I run the simulation in MMA10 or in the browser, the surface in the 3D plot is grayish, whilst in MMA9 or in the CDF Player program it's bluish. That's because (as I remember) MMA10 changed the color schemes of the plots.

Following that last clue I understand that the actual engine operating in the browser is not that of the CDF Player installed on my PC (based on MMA9) but that of MMA10!.

I've also noticed that, when I open my page in the browser, there appears a little icon in the notification area of my OS (Windows 8.1) with the tooltip "Wolfram Mathematica - 10.0.1" (and I had never noticed that icon it before).

Right-clicking on it offers the only option "Quit" and, if I select this option, the embedded CDF in the browser disappears showing just a red exclamation mark.

So my questions are:

  • Why is the notebook ok if run in MMA9 but too slow if run in MMA10?
  • How is that, even if the CDF Player installed in my PC is v.9 (and it executes my cdf file with acceptable speed, like MMA9 does), the Mathematica engine called by my browser accessing my page is some sort of MMA10, and as such is much slower?.
  • Users who have only the CDF Player installed in their PC (without MMA9 or MMA10) would see the simulation in their browser run ok or too slow?

[ I apologize for the length of this post and thank in advance any users that is generous enough to spend some of his/her time to understand the problem and try to give back some of his/her wisdom (answers/solutions/hints) ]

  • 3
    $\begingroup$ It is not necessary to apologize for the length of a well-thought-out question such as this one. $\endgroup$ Jan 6, 2015 at 4:09
  • $\begingroup$ @Luca_M Have you tried running this in the newly released CDFPlayer 10.0.2? $\endgroup$ Feb 12, 2015 at 15:52
  • $\begingroup$ PS its also slow in MMA 10.0.2 $\endgroup$ Feb 12, 2015 at 16:03
  • 1
    $\begingroup$ I have removed the bugs tag. The fact that the performance degrades from one version to the next does not in itself constitute a bug. If you have reported it to Wolfram Research and they have confirmed it's a bug, please update the question with that information. The other way the bugs tag is added is by community consensus, but the comments do not seem to reveal such a consensus. See meta.mathematica.stackexchange.com/q/354. $\endgroup$
    – Michael E2
    Aug 1, 2015 at 2:26


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.