what's wrong with the seagrid spacer?

[leon]

It is said that seagrid can change the space of the grid by change the position of the spacer. I did as the seagrid tutorial, but the spacing is not changing. What's wrong with the seagrid spcaer? Can someone tell me why it doesn't work for me? Thank you very much~

leon

[rsignell]

Leon,

I just tried it, and it works for me. You are supposed to click-and-drag to move the spacers closer or further apart. I had to keep clicking on <toggle>=>refresh to keep the spacers visible, but other than that, it worked okay, and I used it to focus grids toward the center of a test grid as below:



BTW, I just built a new version of the Seagrid Mex files for 2009a. They finally got serious about their threat to take away "finite", and also the mex file code needed all lines with "mxCreateFull" changed to "mxCreateDoubleMatrix". Seagrid is available on our SVN at:

Code: Select all

svn co http://svn1.hosted-projects.com/cmgsoft/m_cmg/trunk/seagrid  seagrid

and the new mex files are in the "mex_matlab78_win32" subdirectory.

-Rich

[leon]

Thank you very much, rich. I know what's the problem. The seagrid that I use is from the RNT toolbox. Maybe it's different or old, so the spacer is not working. I download the one you told me. And the spacer works well. The RNT users should remeber that the RNT seagrid is to some extent different from the latest version. Thank you again,rich.

leon

[hugobastos]

I see this problem a long time ago, but i think u must try to disable the the mex checkbox .

Open Seagrid, Select menu View, and uncheck the Mex box on the end of the window!

Check if works!

[rsignell]

I do NOT recommend avoiding the mex files with SeaGrid. Chuck Denham, the author of this tool, tried to make non-mex versions of the elliptic solver and the other mex files, but it didn't work out so well. If you don't use the mex files, you will get a grid, but it may be wildly nonorthogonal.

[leon]

Actually, when I　use the seagrid in RNT toolbox to compute the land mask, it always costs such a long time. Why does it cost such a long time? I know it depends on the number of grids and the computing domain. Are there other reasons? Which process costs the longest time? Can we improve it ?

[rsignell]

Is it really the land masking that is taking a long time, or the bathymetry gridding?

Seagrid uses a rather slow inverse distance algorithm for bathymetry interpolation, so if you have a large number of bathymetry points, it takes a long time.

There is really no need to do bathymetry gridding in Seagrid: there is an argument that it really should be done by a separate program as a 2nd step. Yet it's nice to see some representation of the bathymetry when you are fiddling with the grid. So I often subsample the bathymetry points so that gridding in Seagrid is quick, then after the file is saved, redo the bathymetry interpolation outside of Seagrid using the full set of bathymetric points or another method (interp2 from a high res DEM, etc.)

[leon]

Yes,it's the land mask computing which cousts a long time. But the seagrid I use is from the RNT toolbox( http://eros.eas.gatech.edu/RNT/ ). Maybe there are some different. I usually load the coastline first, then setup the grids and computing land mask. Finally, adding bathyment and editing land mask.

leon

[hugobastos]

I do NOT recommend avoiding the mex files with SeaGrid. Chuck Denham, the author of this tool, tried to make non-mex versions of the elliptic solver and the other mex files, but it didn't work out so well. If you don't use the mex files, you will get a grid, but it may be wildly nonorthogonal.

Nice to know. Thankz!

[nacholibre]

I do NOT recommend avoiding the mex files with SeaGrid. Chuck Denham, the author of this tool, tried to make non-mex versions of the elliptic solver and the other mex files, but it didn't work out so well. If you don't use the mex files, you will get a grid, but it may be wildly nonorthogonal.

You may find an exaggerated example of a grid generated with/without mex files. It is most likely that the grid generated without the mex files will be geographically incorrect in the end.

[stevekwan]

is it possible to get spacers to appear on all 4 sides? whenever i make change the spaces from sides 1 and 3 to 2 and 4, my grid gets all skewed. is there any way to stabilize the grid spacings before I change sides?

[kate]

This method doesn't let you have spacers on all four sides.

[stone]

Hi,
I download seagrid from http://svn1.hosted-projects.com/cmgsoft ... nk/seagrid
When I use spacing ,It does the spacing work,But Matlab always crash immediately with a error "terminate called after throwing an instance of 'Mathworks::System::SimpleException' "
My OS is Ubuntu i686 with Matlab 2008a
Mathworks support on http://www.mathworks.com/support/soluti ... n=1-36MHWQ does not work either.
What is wrong there?