[ale35]

Dear Ronald,

adding to my previous post, I would like to ask your advice on a parallel hashing issue which might be connected.

I am currently generating a set of parametrized cells using a function. Here I create an interconnect object using my parameters and then I draw some geometries using it. I have noticed that even if all the parameters are the same, calling the Interconnect() command always generates a new object. This means that even if I generate identical cells, they will still be separate.

Is there a way to check whether a specific interconnect object already exists?

An example code:

@nd.bbu.hashme('wg')
def wgs(WG_width):
    with nd.Cell(name = inspect.currentframe().f_code.co_name) as wg:
        ic = Interconnect(width = WG_width, layer = xx)
        ic.taper(...).put()
    return wg

Here if I call wgs() multiple times passing the same WG_width, I will end up with multiple identical cells, each with a different (but identical) interconnect object.

Thank you once again for your help.

[ale35]

Thank you Ronald,

using the hashme decorator solved the warning issue indeed. However, now I get some messages like:

“ERROR: bb_util.py: Reusing a basename across functions: ‘wg’ ”

Am I not using this correctly?

 

 

[ale35]

Thank you for your answer Ronald.

Unfortunately ticking the ‘Remove all variable before execution’ box does not solve the issue, and neither I could solve it with magic commands. Deleting the Nazca arrays frees a bit of memory, but leaves the majority accumulating.

Indeed I reuse the same iPython console, i.e. I just rerun my script in Spyder. the only way I can free my RAM is by manually closing the console every few run.

[Author]

Posts