For running on supercomputers that do not support Python, and for potential speed gains, it would be useful to be able to export a PyNN script as the equivalent sli script.
For scripts that do not involve querying the NEST kernel, this can be most easily done in PyNEST by replacing the sli_push and sli_run functions - Jochen plans to work on this.
For scripts that would involve use of sli_pop we will need to use a proxy backend that does the necessary book-keeping. Such a proxy is anyway needed elsewhere, for MUSIC (see ticket:205), for NineML/NeuroML export and for testing.
Imported from Trac ticket:212
Opened: 2011-11-29 14:26:19
Last modified: 2011-11-29 14:26:19
Component: nest
Priority: minor
Owner: apdavison
Reporter: apdavison
Most usage of NEST appears to be via PyNEST now, so I'm closing this because we have many other priorities. If someone really wants this feature, a pull request would be welcomed!
For running on supercomputers that do not support Python, and for potential speed gains, it would be useful to be able to export a PyNN script as the equivalent sli script.
For scripts that do not involve querying the NEST kernel, this can be most easily done in PyNEST by replacing the
sli_push
andsli_run
functions - Jochen plans to work on this.For scripts that would involve use of
sli_pop
we will need to use a proxy backend that does the necessary book-keeping. Such a proxy is anyway needed elsewhere, for MUSIC (see ticket:205), for NineML/NeuroML export and for testing.Imported from Trac ticket:212
Opened: 2011-11-29 14:26:19 Last modified: 2011-11-29 14:26:19 Component: nest Priority: minor Owner: apdavison Reporter: apdavison