crashing Mathcad...

[Efried]

Since I get quite some of those strange error message before loosing my work (no I alreday know that I have to store in as *.mcd before)

I share the screens with you.

[RichardJ]

Looks a bit like a vapor phase spectrum. Is it?

[Efried]

no actually its a simulation of a solar collector + tank system over 180 days 😉

[VladimirN]

What a version of Mathcad?

[Efried]

15M010

[VladimirN]

And Windows OS?

[Efried]

May be also memory overrun is contributing - apart from avoiding to fill long vectors (500,000+ items) what is the best strategy to reduce the load on mathcad? I already know that there is no way to declare variables as short, also there are no structures which could help to reduce memory demand , but may be I missed something.

One of the reasons is that an programm which should run one hour if you extrapolate the time measurements from a test run, runs endlessly when upscaled.

[Efried]

translating: mathcad has found an user defined stop-point (pause), aborting....

[VladimirN]

Can you attach this Mathcad worksheet?

[ValeryOchkov]

Not welcome to - http://communities.ptc.com/thread/37268

[Efried]

Hi, attached a new error for your entertainment, but the last days only memory overrun prevents from saving in xml format - may be the number of points in plots is to high...

[VladimirN]

Without your Mathcad worksheet we cannot say anything.

[Efried]

Vladimir, I'm not charging you with programming support, but posing general questions to PTC like:

- what are the limits for the amount of points per plot?

- does PTC offer explanations for the mentioned error messages?

- does PTC offer a more robust xml- generator for exchange with the defaulting one?

- future versions may introduce memory saving variable constructs?

[RichardJ]

- what are the limits for the amount of points per plot?

500000. Then it complains, but it does not crash.

- future versions may introduce memory saving variable constructs?

Not likely, but Prime 2.0 has full 64 bit support.

[Efried]

Isn't it silly to have 500,000 as limit whereas a year like this has 525,600 minutes? So you avoid a full year simulation.

It is correct that it does not crash immediately when working in the sheet having 500,000 points, but you can't save as xmcd if you don't disable saving images in the XML options. Finally, yes it does crash frequently anyhow, so exceeding the limits may contribute...

[RichardJ]

Isn't it silly to have 500,000 as limit whereas a year like this has 525,600 minutes? So you avoid a full year simulation.

When someone chose an arbitrary limit of 500,000 I doubt they spent time trying to figure out all the possible ways you could get a number slightly larger than that.

Why do you want to plot 525,000 points anyway? There is no device you could put that graph on where you could resolve that many points. So if it's causing you problems, just decimate the array to 52500 points and plot that. I actually find it hard to imagine why you need a resolution of 1 minute in a 1 year simulation in the first place, but you certainly don't need it for a plot.

[VladimirN]

You can ask also all these questions to PTC's technical support.

[Efried]

I just found a thread about memory issues here

http://communities.ptc.com/message/155503

allow me to state that I have quite some vectors having a lenght of 525,599.

Is there a way to check the risk if Mathcad runs against limits?

Memory demand increased from 124,168 to 266,452 during the calculation, the page faults soared however...

[Efried]

What helpped for the moment to avoid GDI+ and out of memory problems when saving the xmcd file is to disable storing images in the XML-settings. Hoewever Mathcad sometimes behaves strangely after having caclulated and displayed 500kPoints and I have to restart if the results are doubtful. This does not create trust in an engineering product...