Better type stability #247
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
bvdmitri
reviewed
Jul 16, 2024
| plugins = filter(type, getplugins(model)) | ||
| return foldl(plugins; init = (label, nodedata)) do (label, nodedata), plugin | ||
| return preprocess_plugin(plugin, model, context, label, nodedata, options) | ||
| end | ||
| end::Tuple{NodeLabel, NodeData} |
There was a problem hiding this comment.
Suggested change
| end::Tuple{NodeLabel, NodeData} | |
| end |
|
|
||
| Base.show(io::IO, label::NodeLabel) = print(io, label.name, "_", label.global_counter) | ||
| Base.:(==)(label1::NodeLabel, label2::NodeLabel) = label1.name == label2.name && label1.global_counter == label2.global_counter | ||
| Base.hash(label::NodeLabel, h::UInt) = hash(label.name, hash(label.global_counter, h)) | ||
| Base.hash(label::NodeLabel, h::UInt) = hash(label.global_counter, h) |
There was a problem hiding this comment.
IMO it should actually be fine, hash can collide in principle. Also the documentation says
hash(x[, h::UInt]) -> UInt
Compute an integer hash code such that isequal(x,y) implies
hash(x)==hash(y). The optional second argument h is another hash code to be
mixed with the result.
So in principle our a == b implies hash(a) == hash(b) and it follows the specification as documented. The opposite relation is not necessarily required.
bvdmitri
reviewed
Jul 16, 2024
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #247 +/- ##
=======================================
Coverage 90.85% 90.86%
=======================================
Files 15 15
Lines 2088 2090 +2
=======================================
+ Hits 1897 1899 +2
Misses 191 191 ☔ View full report in Codecov by Sentry. |
Co-authored-by: Bagaev Dmitry <[email protected]>
Co-authored-by: Bagaev Dmitry <[email protected]>
Co-authored-by: Bagaev Dmitry <[email protected]>
bvdmitri
approved these changes
Jul 16, 2024
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.
Hash now depends on only the global counter; we use a type-stable version of
to_symbolto hash constant nodes to the context.FactorIDis parameterized because the type instability of having them in the dictionary is better than the type instability of the hashing and comparison downstream from adding them to the dictionary. The plugin system has a type assertion (which acts as a type hint to the compiler) to make the variable node creation type stable. This is ugly but I couldn't get the compiler to understand that this function always returns aTuple{NodeLabel, NodeData}.Everything considered this is a +-25% speedup of the entire model creation, with around 25% less memory consumption