Speedup Big FromListConnector #1095
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
andrewgait
approved these changes
Jun 24, 2021
| self.splitter = MockSplitter(slices) | ||
|
|
||
|
|
||
| class MockMachieVertex(object): |
There was a problem hiding this comment.
MockMachieVertex -> MockMachineVertex
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Depends on:
SpiNNakerManchester/PACMAN#398
tested by:
SpiNNakerManchester/IntegrationTests#64
BUG FIX: SynapseIORowBased
Now asks splitters for the slices!
Method could_connect changed to use vertexes
This allows __split_conn_list to be used!
__split_conn_list now maps only pairs of slices that connect
Instead of checking if not none and then length of item can now just check in
_split_connections
always sets _split_connections to a dict even if empty
Only maps if len(indices) > 0
exit because slices seen is most common so moved first
create_synaptic_block
exit fast if not in __split_conn_list which includes the not len(self.__sources) case
tests show partitioner time cut by over 80 in large fromLists
for example:
import numpy
import spynnaker8 as sim
n_sources = 100000
n_targets = 100000
n_connections = 10000000
sources = numpy.random.randint(0, n_sources, n_connections)
targets = numpy.random.randint(0, n_targets, n_connections)
connection_list = numpy.dstack((sources, targets))[0]
sim.setup(timestep=1.0)
sim.set_number_of_neurons_per_core(sim.IF_curr_exp, 100)
pop_1 = sim.Population(n_sources, sim.IF_curr_exp(), label="pop_1")
input = sim.Population(n_targets, sim.SpikeSourceArray(spike_times=[0]), label="input")
input_proj = sim.Projection(input, pop_1, sim.FromListConnector(connection_list),
synapse_type=sim.StaticSynapse(weight=5, delay=1))
simtime = 10
sim.run(simtime)
In Master:
0:04:43.446246 taken by SpYNNakerSplitterPartitioner
This PR:
0:00:38.930407 taken by SpYNNakerSplitterPartitioner