Test diode_chamber example - #1
Conversation
|
|
||
| Bugs: | ||
| ==== | ||
| 1. Using same kernel seeds causes different behavior: the position are generated randomly and setting Python random.seed does not solve the problem. |
There was a problem hiding this comment.
Yeah, you need to use numpy.random.seed
There was a problem hiding this comment.
oh. :/ I tried numpy.seed
... but this happens when you re not coding in Python for 1 year
| ==== | ||
| 1. Using same kernel seeds causes different behavior: the position are generated randomly and setting Python random.seed does not solve the problem. | ||
|
|
||
| 2. The simulation sometimes (not reproducible because it depends on the random generatorr) stuck in this error. It happens more often within long simulations (4 days+) |
There was a problem hiding this comment.
With the latest version of examples, it should not get stuck but should raise an error and then close.
Was the code really stalled? I know there is one remaining error that sometimes occurs when a branch is retracting, is that it?
There was a problem hiding this comment.
I guess so, because maybe it was happening only with resource based elongation. I did not connected the two things but indeed the models I evaluate are all NOT res_...
I will check it.
| 2. The simulation sometimes (not reproducible because it depends on the random generatorr) stuck in this error. It happens more often within long simulations (4 days+) | ||
| 'src/kernel/space_manager.cpp:310 | ||
|
|
||
| 3. Saving file to graph with 'graph.to_file("diode.el")' gives an error [1]. In the last simulations it is not happening anymore. It may be due to some specific gc configuration. |
There was a problem hiding this comment.
can you give me the full error message so I know where this is happening?
There was a problem hiding this comment.
if you run the file it is fine for you the graph.to_file? It may be due to my installation of NGNT, it may be old.
| 3. Saving file to graph with 'graph.to_file("diode.el")' gives an error [1]. In the last simulations it is not happening anymore. It may be due to some specific gc configuration. | ||
| TypeError: int() argument must be a string, a bytes-like object or a number, not 'NoneType' | ||
|
|
||
| 4. It always better to close the Ipython IDE than do the ds.reset_kernel() only. In the latter case sometimes it seg_fault |
There was a problem hiding this comment.
I never encountered that... could you launch it in gdb and track the segfault if this happens?
2 participants
In this PR:
Refined the example and made comments in the Readme (within diode_chambers folder)