By using the neuromorphic hardware, we need to set in the setup high-level function parameters like useSystemSim or hardware to quote the most used. Typically, we use:
from pyNN.hardware.brainscales import *
setup(loglevel=2,useSystemSim=True, hardware=hardwareSetup['one-hicann'],
timestep=0.1,min_delay=0.1,max_delay=4.0)
However, these parameters (useSystemSim, hardware) have no meaning for the other simulators. We would expect pyNN to print a WARNING if these parameters are used with other simulators (nest, brian, etc...). However, in the current state of the software, it throws an exception and exits.
I think this should be changed to a WARNING (as indicated in the paragraph above). Indeed, in the current state of the hardware, we need to make simulator-specific setup. It means to make simulator-specific high-level pyNN file, which goes against the philosophy of pyNN
it is the use of the hardware.brainscales specific function hardwareSetup that brings the error, not the fact that the keywords are not used by other backends.
So, it is not an issue for pyNN.
By using the neuromorphic hardware, we need to set in the setup high-level function parameters like useSystemSim or hardware to quote the most used. Typically, we use:
from pyNN.hardware.brainscales import * setup(loglevel=2,useSystemSim=True, hardware=hardwareSetup['one-hicann'], timestep=0.1,min_delay=0.1,max_delay=4.0)
However, these parameters (useSystemSim, hardware) have no meaning for the other simulators. We would expect pyNN to print a WARNING if these parameters are used with other simulators (nest, brian, etc...). However, in the current state of the software, it throws an exception and exits.
I think this should be changed to a WARNING (as indicated in the paragraph above). Indeed, in the current state of the hardware, we need to make simulator-specific setup. It means to make simulator-specific high-level pyNN file, which goes against the philosophy of pyNN