spinn_front_end_common.mapping_algorithms.on_chip_router_table_compression package¶
Submodules¶
spinn_front_end_common.mapping_algorithms.on_chip_router_table_compression.compression module¶
-
spinn_front_end_common.mapping_algorithms.on_chip_router_table_compression.compression.
mundy_on_chip_router_compression
(routing_tables, transceiver, machine, app_id, system_provenance_folder, compress_only_when_needed=True, compress_as_much_as_possible=False)[source]¶ Load routing tables and compress then using Mundy’s algorithm
Parameters: - routing_tables – the memory routing tables to be compressed
- transceiver – the spinnman interface
- machine – the SpiNNaker machine representation
- app_id – the application ID used by the main application
- system_provenance_folder – the path to where to write the data
- compress_as_much_as_possible (bool) – If False, the compressor will only reduce the table until it fits in the router space, otherwise it will try to reduce until it until it can’t reduce it any more
- compress_only_when_needed (bool) – If True, the compressor will only compress if the table doesn’t fit in the current router space, otherwise it will just load the table
Returns:
-
spinn_front_end_common.mapping_algorithms.on_chip_router_table_compression.compression.
pair_compression
(routing_tables, transceiver, executable_finder, machine, app_id, provenance_file_path, compress_only_when_needed=False, compress_as_much_as_possible=True)[source]¶ Load routing tables and compress then using Pair Algorithm
- See pacman/operations/router_compressors/pair_compressor.py which is the
- exact same algorithm implemented in python
Parameters: - routing_tables – the memory routing tables to be compressed
- transceiver – the spinnman interface
- executable_finder –
- machine – the SpiNNaker machine representation
- app_id – the application ID used by the main application
- provenance_file_path – the path to where to write the data
- compress_as_much_as_possible (bool) – If False, the compressor will only reduce the table until it fits in the router space, otherwise it will try to reduce until it until it can’t reduce it any more
- compress_only_when_needed (bool) – If True, the compressor will only compress if the table doesn’t fit in the current router space, otherwise it will just load the table