| 3 |
Below is a list of proposed, self guided exercises. I generally tried to order |
Below is a list of proposed, self guided exercises. I generally tried to order |
| 4 |
the exercises by increasing complexity. While none of them is really challenging, |
the exercises by increasing complexity. While none of them is really challenging, |
| 5 |
the various exercises aim to give you with first hand experience with the data sets |
the various exercises aim to give you with first hand experience with the data sets |
| 6 |
and tools discussed over the course of the IAP activity. |
and tools discussed over the course of the IAP activity. Here it is assumed that you have |
| 7 |
|
completed the set up of the activities as explained in guidelines/iap-idma-instructions. |
| 8 |
|
|
| 9 |
tips : - look for answers/examples in the programs we ran together in class #1 and #2 |
tips : - look for answers/examples in the programs we ran together in class #1 and #2 |
| 10 |
- type ‘help read_nctiles’ in matlab and similarly for all other functions |
- type ‘help read_nctiles’ in matlab and similarly for all other functions |
| 11 |
- use the matlab debugger to go through computations step by step |
- use the matlab debugger to go through computations step by step |
| 12 |
|
|
|
notes on matlab software and exercises: |
|
|
—————————————————— |
|
|
|
|
|
Having the up-to-date matlab software (gcmfaces and MITprof) set-up is pre-requisite. |
|
|
|
|
|
1) if you did this set-up by following steps 1 and 2 of computing/iap-idma-readme (i.e. |
|
|
using setup_gcmfaces_and_mitprof.csh) then the ‘software exercise’ #1 is for you |
|
|
|
|
|
2) If you operate on a windows PC where shell scripting, cvs, etc cannot be relied upon, |
|
|
then ‘software exercise’ #2 below is for you. |
|
|
|
|
|
software exercises: |
|
|
————————— |
|
|
|
|
|
1) in a terminal window, go to your copy of gcmfaces and update it using cvs, |
|
|
then do the same thing with MITprof. |
|
|
tip : see http://mitgcm.org/public/using_cvs.html |
|
|
|
|
|
2) set-up the matlab software manually |
|
|
tip : see instructions provided at the beginning of |
|
|
http://mitgcm.org/viewvc/*checkout*/MITgcm/MITgcm_contrib/gael/matlab_class/gcmfaces.pdf |
|
|
|
|
| 13 |
Argo profile data exercises: |
Argo profile data exercises: |
| 14 |
———————————— |
———————————— |
| 15 |
|
|
| 67 |
————————— |
————————— |
| 68 |
|
|
| 69 |
1) increase experiment duration for adjustment.cs-32x32x1, |
1) increase experiment duration for adjustment.cs-32x32x1, |
| 70 |
tutorial_held_suarez_cs, and tutorial_plume_on_slope |
tutorial_held_suarez_cs, and tutorial_plume_on_slope |
| 71 |
|
by editing their respective input/data as follows: |
| 72 |
tips : - if using testreport for longer runs then use the -fast option |
- tutorial_held_suarez_cs/input/data: |
|
- for tutorial_held_suarez_cs/input/data: |
|
| 73 |
< nTimeSteps=16, |
< nTimeSteps=16, |
| 74 |
> nTimeSteps=69120, |
> nTimeSteps=69120, |
| 75 |
- for tutorial_plume_on_slope/input/data: |
- tutorial_plume_on_slope/input/data: |
| 76 |
< nTimeSteps=20, |
< nTimeSteps=20, |
| 77 |
> nTimeSteps=8640, |
> nTimeSteps=8640, |
| 78 |
- for adjustment.cs-32x32x1/input/data: |
- adjustment.cs-32x32x1/input/data: |
| 79 |
< nTimeSteps=24, |
< nTimeSteps=24, |
| 80 |
> endTime=172800., |
> endTime=172800., |
| 81 |
|
|
| 82 |
2) compile, run, and display utorial_held_suarez_cs with diagnostic package |
2) compile and run the experiments using testreport using the command |
| 83 |
|
./testreport -fast -t tutorial_plume_on_slope |
| 84 |
|
and accordingly for the other two. Note: the -fast option activates |
| 85 |
|
compiler optimization, which accelerates the runs greatly. |
| 86 |
|
|
| 87 |
|
notes : - to learn more about testreport type 'testreport --help' or |
| 88 |
|
see mitgcm.org/public/devel_HOWTO/devel_HOWTO.pdf |
| 89 |
|
- 'testreport -clean' is useful to reset an experiment. |
| 90 |
|
- testreport links the various inputs in the 'run/' subdirectory |
| 91 |
|
and then automatically runs the model ('mitgcmuv'). As long as |
| 92 |
|
you do not empty 'run/' then you can exectute the model directly |
| 93 |
|
using this command: './mitgcmuv > output.txt'. |
| 94 |
|
|
| 95 |
|
4) use the mitgcm_plot_adju_cs32.m, mitgcm_plot_held_suarez.m, and mitgcm_plot_plume.m |
| 96 |
|
(http://mitgcm.org/viewvc/MITgcm/MITgcm_contrib/gael/comm/course-idma2016/matlab/) |
| 97 |
|
to display results (type e.g. 'help mitgcm_plot_adju_cs32' first). |
| 98 |
|
|
| 99 |
tips : - adjustment.cs-32x32x1 provides a working example using pkg/diagnostics |
notes: this assumes that gcmfaces has been installed and added to the Matlab |
| 100 |
- pkg/diagnostics needs to be activated at compile time (packages.conf), and |
path as explained at the beginning of idma2016-instructions.pdf |
|
at run time (data.pkg), and then configured at run time (data.diagnostics) |
|
| 101 |
|
|
Parent Directory
| 
Revision Log
| 
Revision Graph
| 
Patch