2001i worksheets in MC 14

[ptc-3798987]

Large sets of differential equations, which can be solved numerically by MC 2001i, can not be processed by MC 14. In MC 14 worksheets of 2001i are blown up to about 8-fold volume, they are correctly computed, but can not be saved, if a certain volume is overstepped. Three 2001i worksheets of different volume are attached:





1. 194 KB (solved and saved in MC 14)





2. 429 KB (solved and saved in MC 14)





3. 526 KB (solved but not saved i9n MC 14)





Why become 2001i worksheets enlarged in MC 14? Is this an error or the result of a new program structure?

[IRstuff]

M14 file format is text-based, like it was in the original versions of Mathcad. However, the text format is per XML syntax, which is even larger that MC's original purpose-built text file format.

The good news is that you can save your M14 files in the zipped (compressed) format, and the files will compress nicely.

Nonetheless, since the 2001i file format is binary, there will still be considerable bloat in M14 compared to 2001i.

TTFN,
Eden

[TomGutman]

Save in compressed format (XMCDZ) after setting the file properties not to save generated images. That should result in reasonably small files.
__________________
� � � � Tom Gutman

[RichardJ]

On 11/30/2009 2:50:51 PM, Tom_Gutman wrote:
>after setting the file
>properties not to save
>generated images.

That is the key. If that it is set to save images I get an out of memory error when trying to save the third file.

Richard

[TomGutman]

I was able to save that third file in XMCD format. But it is very large. Much more effective is the attached XMCDZ format, smaller than the 2001i format.

I do notice that you have a lot of calculations of large vectors. That is rather hard on the memory management (which is rather poor in MC14 anyway), as each vector result has to be stored as a complete vector. I suggest using function definitions instead, and only evaluating the vectors when you need the actual values for plotting or other display.
__________________
� � � � Tom Gutman

[ptc-1368288]

Comments for comments:

1. Hard to explain that such a chain of functions/solution is not in ladder style, you have them spreading over 3 pages in all sorts of soup.
2. I don't understand 360000*50 ?
Isn't same as 1.8*10^7 ?
3. The model is not supported neither sourced.
4. Too much work to cleanup.
5. It looks like you could carry lot of the work sheet in scalar mode and vectorize at the point of use. When you vectorize too early, it creates a vector/matrix of values that finally at 20000 elements each DO CONSUME a lot of memory. It could very well be that vectorizing is only needed at the reciprocal. All those vector/matrix vectorized values can be collected in a program, which program you disable before saving and save in probably < 25 KB.



How come a work sheet can be so messy ? First time a 14 saved lower is so much "pinged-ponged" lottery style ! Few collapsed areas and the work sheet would be 2 to 3 pages. Despite reported problems with 14, I have serious doubts that it can't be done correctly and saved readable... 14 done correctly and saved readable were my two points.

I have just read once down the first work sheet.
Quite a scare crow ! Is it too cruel ?

jmG