[Py] Improve embeddings #12
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
We want to allow for more chain scenarios, e.g. add a random chain where both source and target node are in chains themselves. This led to overcomplicated code having to deal with chains that are placed on edges instead of nodes. Thus, this commit is a working example for better chain placement (although not perfect), but definitely needs to be redone in the next commits (e.g. using the already-used mapping from H to G instead of essentially redundant chains on edges).
This only affects the embedding and solver module.
Beforehand, we encoded supernodes using edge costs, e.g. edge with cost 0 was no chain, an edge with cost 1 indicated that the two nodes mapped to the same node_H, thus they belong to the same supernode. However, this was essentially redundant as we also stored the mapping between input graph H and hardware graph G in the mapping class. We leverage this class now and encode supernodes directly, thus getting rid of edge chains (remnants are still there, will clean up later). It might be worth to keep the term "chain" to indicate supernodes that consist of more than one node. This change makes the embedding itself worse, e.g. K4 does not work anymore. This is since we now also allow to merge random nodes to supernodes where both source and target are in chains. We need to implement the evolutionary strategy where we select the mutation that yields the best result, in our case where the most edges can be embedded (to achieve a "real" local maximum). Also changed: - Changed lists to sets in various places - Adapted drawing class to the new changes (to properly draw supernodes) - Made logging output more precise
Split the method into multiple pieces for better readability and fixed several bugs that occurred due to wrong order of execution. It is still quirky that we can't embed edges freely since all nodes on that edge must have supernode (in the current implementation). That's why there are some special constructs, e.g. we need to discard nodes from sets to make sure not the change results after having embedded them in supernodes. There is room for improvement here. Fixed a really annoying bug: Sets in Python are returned by reference, so we got weird results since we used discard on them assuming we just altered the local variable. Now we copy the set. Fixed another, potential bug where we accidentally added selfloops to the embedding view graph. Also improved error messages.
Note that at this point, we still sometimes use the word "chain", however we don't use chains encoded on edges anymore, but now on nodes themselves. TODO: Unify usage of words "chain" and "supernode" throughout the code.
Also slightly shifted labels of Chimera graph for better readability.
After each generation we also try to remove unnecessary edges.
This way, we don't remove the wrong nodes in our graph that would disconnect the respective subgraph of supernodes (if removed). See https://www.geeksforgeeks.org/articulation-points-or-cut-vertices-in-a-graph/ and https://youtu.be/jFZsDDB0-vo as examples
Previously, we removed all nodes that were not articulation points for every supernode in one go. However, as we remove, the articulation points might change, so we need to recalculate them every time a node is removed.
We don't shift the target node in this strategy.
Also switched to 5x5 Chimera cell
keksklauer4
added a commit
that referenced
this pull request
Apr 9, 2022
* Fix spelling mistake * [Python] Init docs (#9) TL;DR: - Refactor graph python module - Init Sphinx documentation: https://majorminer.readthedocs.io/ * [Py] Implement BFS and DFS for initialization BFS: Breadth-first search DFS: Depth-first search * Refactor undirected Graph and add doc strings. * Init Sphinx ReadTheDocs documentation * Add documentation badge * Add docstrings to graph * Init Chimera graph docs * Add ReadTheDocs configuration to fix doc errors Use Python 3.9 instead of version 3.7 for building. * Add doc strings to embedding graph This partially breaks with some chain logic, so we will need to adjust some functions one abstraction layer above. * Fix get_embedded_nodes * Add overview section to docs including diagrams * Add more docs * Add warning for outdated Readme * [Py] Chimera lattice (#10) * Extend from Chimera cell to whole Chimera lattice - Also outsource test graphs - Do evolution with multiple mutation steps * Simplify connected test graph generation * Try to embed endlessly until embedding found Also save intermediate results as SVG * Add more chain colors and SVG export option * Init loggin instead of prints * Save intermediate SVG graphs * Do not return multiple chains (instead use set) * Add typing to util function * Only use from_nodes which are not in a chain * Tighten error handling * Allow mutation for to_nodes that are in a chain * Include max_total as bound for main_loop passes * Export step-by-step SVG For this, we outsourced the local maximum technique to not be included inside a mutation. It now has to be called from outside (evolution.py). * Change draw direction to horizontal * [Py] Improve embeddings (#12) * ⚠ Overcomplicate logic for random chains We want to allow for more chain scenarios, e.g. add a random chain where both source and target node are in chains themselves. This led to overcomplicated code having to deal with chains that are placed on edges instead of nodes. Thus, this commit is a working example for better chain placement (although not perfect), but definitely needs to be redone in the next commits (e.g. using the already-used mapping from H to G instead of essentially redundant chains on edges). * Rename from_node->source, to_node->target This only affects the embedding and solver module. * Shift from edge to node supernode encoding Beforehand, we encoded supernodes using edge costs, e.g. edge with cost 0 was no chain, an edge with cost 1 indicated that the two nodes mapped to the same node_H, thus they belong to the same supernode. However, this was essentially redundant as we also stored the mapping between input graph H and hardware graph G in the mapping class. We leverage this class now and encode supernodes directly, thus getting rid of edge chains (remnants are still there, will clean up later). It might be worth to keep the term "chain" to indicate supernodes that consist of more than one node. This change makes the embedding itself worse, e.g. K4 does not work anymore. This is since we now also allow to merge random nodes to supernodes where both source and target are in chains. We need to implement the evolutionary strategy where we select the mutation that yields the best result, in our case where the most edges can be embedded (to achieve a "real" local maximum). Also changed: - Changed lists to sets in various places - Adapted drawing class to the new changes (to properly draw supernodes) - Made logging output more precise * Rewrite extend_random_supernode and fix bugs Split the method into multiple pieces for better readability and fixed several bugs that occurred due to wrong order of execution. It is still quirky that we can't embed edges freely since all nodes on that edge must have supernode (in the current implementation). That's why there are some special constructs, e.g. we need to discard nodes from sets to make sure not the change results after having embedded them in supernodes. There is room for improvement here. Fixed a really annoying bug: Sets in Python are returned by reference, so we got weird results since we used discard on them assuming we just altered the local variable. Now we copy the set. Fixed another, potential bug where we accidentally added selfloops to the embedding view graph. Also improved error messages. * Get rid of chain remnants Note that at this point, we still sometimes use the word "chain", however we don't use chains encoded on edges anymore, but now on nodes themselves. TODO: Unify usage of words "chain" and "supernode" throughout the code. * Explicitly use sets instead of lists * Draw supernode colors on nodes (not only on edges) Also slightly shifted labels of Chimera graph for better readability. * Change color brightness instead of transparency * Fix bug not operating on playground embedding * Fix bug forgot removing target from prev supernode * Init basic hill climbing (evolution) After each generation we also try to remove unnecessary edges. * Move supernode connectiveness check to embedding * Remove redundant nodes probabilistically * Implement articulation point algorithm (cut node) This way, we don't remove the wrong nodes in our graph that would disconnect the respective subgraph of supernodes (if removed). See https://www.geeksforgeeks.org/articulation-points-or-cut-vertices-in-a-graph/ and https://youtu.be/jFZsDDB0-vo as examples * Fix removal of redundant nodes Previously, we removed all nodes that were not articulation points for every supernode in one go. However, as we remove, the articulation points might change, so we need to recalculate them every time a node is removed. * Add strategy to extend supernode without shifting We don't shift the target node in this strategy. * Fix articulation point (don't include removed nodes) * Remove redundant nodes as last resort in the end Also switched to 5x5 Chimera grid At this point K6 embedding on a 3x3 cell is working. Tried K12 on a 5x5 Chimera grid, which is not yet yielding good valid embeddings. TODO: Code cleanup, especially the embedding solver that got kind of cluttered. * [Cpp] Refactoring and mutation class. (#11) * [Cpp] Implemented visualizer for generic graphs. * [Cpp] Started reworking parallelization concept for iterative local improvement. * [Cpp] Adjusted extend operator. * [Cpp] Continued with mutation manager. * [Cpp] Fundamental rework of the structure * [Cpp] Continued refactoring. * [Cpp] Still refactoring. * [Cpp] Refactoring... * [Cpp] Refactored imports. * [Cpp] Code running again - at least a bit. * [Cpp] Continued fixing problems * [Cpp] Refactoring. * [Cpp] Refactored embedding state and added generic iteration methods. * [Cpp] Some more loop replacements * [Cpp] Fixing shifting operator. * [Cpp] Implemented random gen and started working on shifting. * [Cpp] Working on shifting. * [Cpp] Fixed mutations. * [Cpp] Started implementing annealing-based super vertex reducer. * [Cpp] Continuing super vertex reducer. * [Cpp] Fixed embedding invalidating bug. * [Cpp] Working on extend. * [Cpp] Implemented reducer as mutation. * [Cpp] Started implementing evolutionary csc reducer. * [Cpp] Continuing csc reducer. * [Cpp] Added K15 for testing. * [Cpp] Implemented evolutionary CSC reducer. * [Cpp] Fixed some csc evo bugs. * [Cpp] Added super vertex replacer. * [Cpp] Fixed bugs. * Update README.md Co-authored-by: Splines <37160523+Splines@users.noreply.github.com>
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.
No description provided.