GSvit documentation

open source FDTD solver with GPU support

User Tools

Site Tools


start:first_simulation

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
start:first_simulation [2018/09/04 16:00]
pgrolich
start:first_simulation [2019/01/14 07:58] (current)
pgrolich
Line 27: Line 27:
 {{:​start:​screenshots:​first-simulation-init.png?​400|}} {{:​start:​screenshots:​first-simulation-init.png?​400|}}
  
-It is reasonable to save your parameter file at this stage using <i>File->​Save</i> command. We can do it also later, just before running computation,​ however doing it right now will set the current directory and other input/​output files set later can be referred to it (instead of using absolute path). Moreover, by saving parameter file we also set default directory for temporary files (e.g. generated sources), which is also important.+It is reasonable to save your parameter file at this stage using //File->​Save/​command. We can do it also later, just before running computation,​ however doing it right now will set the current directory and other input/​output files set later can be referred to it (instead of using absolute path). Moreover, by saving parameter file we also set default directory for temporary files (e.g. generated sources), which is also important.
  
 First we will add **point source**. This can be done using menu entry //Edit parameters->​Add point source//. First we will add **point source**. This can be done using menu entry //Edit parameters->​Add point source//.
Line 41: Line 41:
 file much faster and see the position changes immediately.//​ file much faster and see the position changes immediately.//​
  
-Next we will add <b>point and image outputs</b>. This is done by menu entry //Edit parameters->​Add point output// and //Edit parameters->​Add image output// that again open dialogues for specifying how often (parameter "​skip"​) and which component should be output. Point output is collected from a single point within computational volume and is stored into a text file. Image output comes to general output file that can be opened by [[http://​gwyddion.net|Gwyddion]] open source software. A cross section in certain normal direction is output, selected as shown below. We will add a cross section in every direction, so we will call the dialog three times.+Next we will add **point and image outputs**. This is done by menu entry //Edit parameters->​Add point output// and //Edit parameters->​Add image output// that again open dialogues for specifying how often (parameter "​skip"​) and which component should be output. Point output is collected from a single point within computational volume and is stored into a text file. Image output comes to general output file that can be opened by [[http://​gwyddion.net|Gwyddion]] open source software. A cross section in certain normal direction is output, selected as shown below. We will add a cross section in every direction, so we will call the dialog three times.
  
 {{:​start:​screenshots:​first-simulation-image-output.png?​400|}} {{:​start:​screenshots:​first-simulation-image-output.png?​400|}}
Line 66: Line 66:
 When computation is finished or killed by the Stop command (again via toolbar or menu), we can **open the output file**. As most of the data are output in Gwyddion file we can directly open it using Gwyddion called from the menu or toolbar. Gwyddion is a powerfull 2D/3D data processing software, however we will only use it to see the images in this example. As image outputs are saved succesively as the GSvit is running, we will be probably having more channels to see in Gwyddion, so we pick what is interesting for us in the Gwyddion Data Browser. When computation is finished or killed by the Stop command (again via toolbar or menu), we can **open the output file**. As most of the data are output in Gwyddion file we can directly open it using Gwyddion called from the menu or toolbar. Gwyddion is a powerfull 2D/3D data processing software, however we will only use it to see the images in this example. As image outputs are saved succesively as the GSvit is running, we will be probably having more channels to see in Gwyddion, so we pick what is interesting for us in the Gwyddion Data Browser.
  
-<a href="​http://​gsvit.net/​images_v18/​ex1_shortwrong.png"><​img src="​http://​gsvit.net/​images_v18/​ex1_shortwrong.png"​ width="​800"></​a>​ +{{:start:first-simulation-shortwrong.png?400|}}
-<​p>​Note that in this example we have clicked on <​i>​Data Process->​Presentation->​Logscale</​i>​ for every displayed output to see the data in +
-better scale. We can see that in different cross-sections of the computational space there is a wave similar to a dipole output.+
  
-<p>If we check the data in further time steps, we can see that they look much worse.</​p>​ +Note that in this example we have clicked on //Data Process->Presentation->​Logscale//​ for every displayed output to see the data in better scale. We can see that in different cross-sections of the computational space there is wave similar to a dipole output.
-<href="​http://​gsvit.net/​images_v18/​ex1_longwrong.png"><​img src="​http://​gsvit.net/​images_v18/​ex1_longwrong.png"​ width="​800"></​a>​+
  
-<​p>​This behavior is caused by unset <​b>​boundary conditions</​b>​ that behave like mirrors here.  +If we check the data in further time steps, we can see that they look much worse.
-In order to let the wave go out of the computational +
-space we need to apply an absorbing boundary conditionso we click on any entry in "​Boundary conditions"​ set to change the boundary conditions +
-on all the faces to a simple absorbing condition (Liao).</​p>​ +
-<img src="​http://​gsvit.net/​images_v18/​ex1_bcs.png">​+
  
-<p>After that (and saving the file, running computation and redisplaying data again) we can see that the output is already propagating +{{:​start:​screenshots:​first-simulation-longwrong.png?​400|}} 
-freely outside of the computation volume.</p> + 
-<a href="​http://​gsvit.net/​images_v18/​ex1_longok.png"><​img src="​http://​gsvit.net/​images_v18/​ex1_longok.png" width="​800"></​a>​+This behavior is caused by unset **boundary conditions** that behave like mirrors here. In order to let the wave go out of the computational space we need to apply an absorbing boundary condition, so we click on any entry in "​Boundary conditions"​ set to change the boundary conditions on all the faces to a simple absorbing condition (Liao). 
 + 
 +{{:​start:​first-simulation-boundary-conditions.png?​400|}} 
 + 
 +After that (and saving the file, running computation and redisplaying data again) we can see that the output is already propagating freely outside of the computation volume. 
 + 
 +{{:start:​screenshots:first-simulation-longok.png?400|}}
  
  
start/first_simulation.1536069603.txt.gz · Last modified: 2018/09/04 16:00 by pgrolich