Source code for pacman.operations.chip_id_allocator_algorithms.malloc_based_chip_id_allocator

# Copyright (c) 2017-2019 The University of Manchester
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.

import logging
from spinn_utilities.progress_bar import ProgressBar
from pacman.exceptions import PacmanConfigurationException
from pacman.model.graphs import (
    AbstractFPGA, AbstractSpiNNakerLink, AbstractVirtual)
from pacman.utilities.algorithm_utilities import (
    machine_algorithm_utilities, ElementAllocatorAlgorithm)

logger = logging.getLogger(__name__)
_LOWER_16_BITS = 0xFFFF








[docs]class MallocBasedChipIdAllocator(ElementAllocatorAlgorithm): """ A Chip ID Allocation Allocator algorithm that keeps track of\ chip IDs and attempts to allocate them as requested """ __slots__ = [ # dict of [virtual chip data] = (x,y) "_virtual_chips" ] def __init__(self): super(MallocBasedChipIdAllocator, self).__init__(0, 2 ** 32) # we only want one virtual chip per 'link' self._virtual_chips = dict() def __call__(self, machine, graph=None): if graph is not None: self.allocate_chip_ids(machine, graph) return machine
[docs] def allocate_chip_ids(self, machine, graph): """ Go through the chips (real and virtual) and allocate keys for each """ progress = ProgressBar( graph.n_vertices + machine.n_chips, "Allocating virtual identifiers") # allocate standard IDs for real chips for x, y in progress.over(machine.chip_coordinates, False): expected_chip_id = (x << 8) + y self._allocate_elements(expected_chip_id, 1) # allocate IDs for virtual chips for vertex in progress.over(graph.vertices): if isinstance(vertex, AbstractVirtual): x, y = self._assign_virtual_chip_info( machine, self._get_link_data(machine, vertex)) vertex.set_virtual_chip_coordinates(x, y)
@staticmethod def _get_link_data(machine, vertex): if isinstance(vertex, AbstractFPGA): link_data = machine.get_fpga_link_with_id( vertex.fpga_id, vertex.fpga_link_id, vertex.board_address) if link_data is None: raise NoFPGALink(vertex) return link_data elif isinstance(vertex, AbstractSpiNNakerLink): link_data = machine.get_spinnaker_link_with_id( vertex.spinnaker_link_id, vertex.board_address) if link_data is None: raise NoSpiNNakerLink(vertex) return link_data else: # Ugh; this means we can't handle link data for arbitrary classes raise PacmanConfigurationException( "Unknown virtual vertex type {}".format(vertex.__class__)) def _assign_virtual_chip_info(self, machine, link_data): # If we've seen the link data before, return the allocated ID we have if link_data in self._virtual_chips: return self._virtual_chips[link_data] # Allocate a new ID and cache it for later chip_id_x, chip_id_y = self._allocate_id() machine_algorithm_utilities.create_virtual_chip( machine, link_data, chip_id_x, chip_id_y) self._virtual_chips[link_data] = (chip_id_x, chip_id_y) return chip_id_x, chip_id_y def _allocate_id(self): """ Allocate a chip ID from the free space """ # can always assume there's at least one element in the free space, # otherwise it will have already been deleted already. free_space_chunk = self._free_space_tracker[0] chip_id = free_space_chunk.start_address self._allocate_elements(chip_id, 1) return (chip_id >> 8), (chip_id & _LOWER_16_BITS)