[release/9.0-staging] Fix to #35239 - EF9: SaveChanges() is significantly slower in .NET9 vs. .NET8 when using .ToJson() Mapping vs. PostgreSQL Legacy POCO mapping #35360
+671
−21
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Outdated suggestions cannot be applied.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.
Fixes #35239
Safer version of #35326
Description
EF9 introduced a change in how we construct ValueComparers for some of our types (specifically collection of scalars/references), in preparation for AOT work. The way the change was implemented may cause a severe performance regression during SaveChanges operation involving multiple entities using collections of primitives (one of our highly requested features).
Customer impact
Customers performing data manipulation operations on entities with collections of primitives may experience significant performance regressions. This may also happen when no data has been changed, but sufficiently large entity graph has been loaded into change tracker. There is no workaround for this issue, apart from changing the model to not use primitive collections (which is unacceptable for majority of customers)
How found
Multiple customer reports on EF 9
Regression
Yes, from EF8. Note: this is a perf regression only, not a functional regression.
Testing
Ad hoc performance test using BenchmarkDotNet. Functional testing already covered by existing tests.
Risk
Low. The patch fix has been limited in scope to reduce the risk. Changes should only affect models with primitive collections. Added quirks just to be sure.
Perf numbers
Over 20x regression between 8 and 9 for the tested scenario.
8.0.11
9.0
9.0.2 (with fix)