KIP-SPIKEY-USERS Archives

2016

KIP-SPIKEY-USERS@LISTSERV.UNI-HEIDELBERG.DE

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Johannes Bill <[log in to unmask]>
Reply To:
Spikey neuromorphic system - User <[log in to unmask]>
Date:
Tue, 10 May 2016 19:06:11 +0200
Content-Type:
text/plain
Parts/Attachments:
text/plain (78 lines)
Hello Michael,

Just as a first guess for approaching your question:

Synapse driver n can only deliver spikes from one out of three 
presynaptic sources: (a) external, (b) neuron n from the left block, or 
(c) neuron (n+192) from the right block; with n=0,..,191.

The options (a), (b) or (c) are chosen upon mapping of the network, i.e. 
during the first call of pynn.run().

Hence, a permutation of neurons in a recurrently connected network could 
lead to the situation that neuron n and (n+192) are supposed to project 
onto the same post-synaptic row (= synapse driver). For instance, it 
would be enough if neuron n projects onto neuron i, 0 < i < 192, and 
neuron (n+192) projects onto neuron j, 0 < j < 192, to raise an exception.

Could this be the case in your network?

Mapping of recurrent networks of size up to 192 is always possible[*].

Best regards,
Johannes

[*] Assuming that short term plasticity is not used.



On 05/10/2016 06:27 PM, Michael Schmuker wrote:
> Hi,
>
> I get a strange error when I use neuronPermutation in my script:
>
> TypeError: Ambiguous synapse driver configuration! Source can't be external and feedback at the same time!
>
> This is thrown in run(), not at setup().
>
> I wasn’t able to reproduce this in a minimal script, it seems that it requires a more complicated network with.
>
> It doesn’t occur when I permute only the first 192 neurons and leave the neurons 193 to 384 unchanged.
>
> Any ideas how to start to debug this? Attached is my spikeyconfig.out.
>
> Thanks in advance!
>
> Best,
>
> Michael
>
>
> ########################################################################
>
> To unsubscribe from the KIP-SPIKEY-USERS list, click the following link:
> http://listserv.uni-heidelberg.de/cgi-bin/wa?TICKET=NzM2MTI0IGJpbGxAS0lQLlVOSS1IRUlERUxCRVJHLkRFIEtJUC1TUElLRVktVVNFUlMgIOT%2FRZ8G2J3K&c=SIGNOFF
>
>
>
>
>
>
> --
> Michael Schmuker
> School of Engineering and Informatics
> University of Sussex
> http://biomachinelearning.net
>
>
> ########################################################################
>
> To unsubscribe from the KIP-SPIKEY-USERS list, click the following link:
> http://listserv.uni-heidelberg.de/cgi-bin/wa?TICKET=NzM2MTI0IGJpbGxAS0lQLlVOSS1IRUlERUxCRVJHLkRFIEtJUC1TUElLRVktVVNFUlMgIOT%2FRZ8G2J3K&c=SIGNOFF
>

########################################################################

To unsubscribe from the KIP-SPIKEY-USERS list, click the following link:
https://listserv.uni-heidelberg.de/cgi-bin/wa?SUBED1=KIP-SPIKEY-USERS

ATOM RSS1 RSS2