Robin Whittle @ on
Error post-processing Meep Rod Dipole example
I carefully ran the Rod Dipole example:
1 – Produce Meep input files only. 2 – Dry-run the current configuration 3 – Fully simulate the current configuration
without any changes or obvious problems. In the relevant session directory (371174L), I had a subdirectory of the name (xxx) I gave my project, and in that directory were three files:
* structure_profile_epsilon.h5
* volume_1_epsilon.h5
* volume_1_norm_epsilon.h5
When I closed that session and tried (session 371221):
4 – Load and post-process a completed simulation
I noticed the system had created a directory “data/sessions/371174L-expired” with all the files (I guess) copied from the “data/sessions/371174L” directory. It then started processing some things, presumably so I could view the fields, according to the new near-a and far-field settings I saw and left unchanged in the “7 Post-Processing” stage.
A bunch of stuff scrolled by and when it finished, the results list was:
* Structural Profile: Dielectric Constant * Material Model: Au_optical – 0.5 to 6.2 eV * Middleware Error Log * Middleware Output Log * Critical Error
In the stuff, were lines including:
Using directory “/home/nanohub/robinwhittle/data/sessions/371174L-expired” as the toolkit output.
The Middware Error Log included:
(I can’t cut and paste from the Java window, this is manually transcribed, and I couldn’t find a file with this in it.)
[Errno 2] No such file or directory: ‘/home/nanohub/robinwhittle/data/sessions/371174L/SaveFiles/Rod Dipole no changes.sim’
Sure enough, there is no such file there – but there is on in the same location in the “expired” directory.
I copied the file from the “expired” location to the “…371174L/SaveFiles/” directory and tried again, closing the 371221 session and starting a new one 371231. This produced even more errors, (again I don’t know where this Middleware Error Log file is in the system, but it appears the software found the file ‘/home/nanohub/robinwhittle/data/sessions/371174L/SaveFiles/Rod Dipole no changes.sim’ inaccessible. I checked via WebDAV and the Storage Manager – and the file was still there, apparently intact.
I am looking forward to using Meep!
– Robin
0 Like 0 Dislike
Todor Antonijevic @ on
Hi, I have similar problem for Rod Dipole example. I am not able to perform the last step “Load and post-process a completed simulation” correctly. Was there any further suggestion/improvements ? Todor
Report abuse Reply
0 Like 0 Dislike
Alexander S McLeod @ on
Report abuse Reply
Robin Whittle @ on
Hi Alex, Thanks for your reply here and for your email. A subsequent attempt to run an example crashed, as you suggested by email, due to the results overrunning my storage allocation. I can’t find any technical documentation on how this Nanohub.org session works – just stuff about the website. In order to make more attempts, I deleted all my files, so I can’t look into log files for a “critical error”. I am attempting to run all three examples, one at a time, without changes, using my full allocation of storage. I have not yet succeeded with “Bowtie Antenna” or “Rod Dipole” and am currently trying an unmodified form of “Bowtie Cross Antenna. This question system strikes me as poorly designed. The composition window is tiny. I can’t update what I wrote. It mangles my carefully formatted text by removing leading white space on lines and by removing some newlines. I think the system should be in fixed width font with no alteration at all of the text I write. This is a technical discussion and we need to be able to post technical text, with white space intact. Meanwhile, my homepage informs me I can “earn 23 points” by closing this question. I don’t see any good purpose for video-game-like “points”. The nanohub.org site seems to be lacking any documentation of the storage capacity each user gets, or the purpose of the various files which are generated for any run of a tool such as PhotonicsTK (Meep). I perceive a mismatch between the sophistication of software such as Meep, together with the power of the Rossmann and Coates Clusters which are needed to run Meep on substantial problems . . . and the very hip Web2.0 appearance of this site with its apparent absence of technical documentation. Its early days for your Meep system, I understand. I thinnk it is an excellent and generous gift to allow people from all over the world to use this program, and the required cluster compute power, for free, with your carefully crafted interface which means they don’t need to write their own .CTL files. I guess I am not the only one having trouble getting the basic examples to work, which for me is a first step before trying to do my own projects. – Robin
Report abuse Reply