Cache the compilation states in compilation order to avoid recalculation #76380
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
Cache the compilation states in compilation order to avoid recalculation GetStatesInCompilationOrder shows up pretty heavily in CPU usage in the typing scenario in the csharp editing speedometer. The enumerating of the ImmutableList and the repeated searching in the ImmutableSortedDictionary are actually more expensive than one would think. GetStatesInCompilationOrder is optimized out of the profile, but it looks like it is accounting for about 5% of CPU. I'm going to run a test insertion with this change and see what the CPU/memory delta of this method looks like with this change.
dotnet-issue-labeler
bot
added
Area-Compilers
untriaged
ToddGrun
commented
Dec 11, 2024
| @@ -397,6 +397,22 @@ public static ImmutableArray<TResult> SelectAsArray<TSource, TResult>(this IRead | |||
| return ImmutableCollectionsMarshal.AsImmutableArray(builder); | |||
| } | |||
|
|
|||
| public static ImmutableArray<TResult> SelectAsArray<TSource, TResult, TArg>(this IReadOnlyCollection<TSource>? source, Func<TSource, TArg, TResult> selector, TArg arg) | |||
| { | |||
| if (source == null) | |||
ToddGrun
commented
Dec 11, 2024
|
VS Test insertion for speedometer results: https://dev.azure.com/devdiv/DevDiv/_git/VS/pullrequest/598012 Run shows similar CPU/Memory in baseline for WithDoNotCreateCreationPolicy With the changes, there is a pretty dramatic difference in CPU usage in WithDoNotCreateCreationPolicy. The ImmutableDictionary/ImmutableList costs in the method are pretty much gone. I do feel like this might have been a lucky run with numbers that were just better, but the WithDoNotCreateCreationPolicy changes are definitely outside the bounds of luck. CPU with changes: Allocations with changes: |
ToddGrun
changed the title
|
@dotnet/roslyn-compiler ptal |
|
@CyrusNajmabadi ptal |
AlekseyTs
approved these changes
Dec 12, 2024
src/Workspaces/Core/Portable/Workspace/Solution/TextDocumentStates.cs
Outdated
Show resolved
Hide resolved
jasonmalinowski
approved these changes
Dec 12, 2024
src/Workspaces/Core/Portable/Workspace/Solution/TextDocumentStates.cs
Outdated
Show resolved
Hide resolved
src/Workspaces/Core/Portable/Workspace/Solution/TextDocumentStates.cs
Outdated
Show resolved
Hide resolved
CyrusNajmabadi
approved these changes
Dec 12, 2024
|
@dotnet/roslyn-compiler -- need 2nd compiler review (fairly trivial change) |
jjonescz
approved these changes
Dec 13, 2024
This was referenced Dec 17, 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.
Cache the compilation states in compilation order to avoid recalculation
GetStatesInCompilationOrder shows up pretty heavily in CPU usage in the typing scenario in the csharp editing speedometer. The enumerating of the ImmutableList and the repeated searching in the ImmutableSortedDictionary are actually more expensive than one would think.
GetStatesInCompilationOrder is optimized out of the profile, but it looks like it is accounting for about 5% of CPU.
I'm going to run a test insertion with this change and see what the CPU/memory delta of this method looks like with this change.
*** CPU in typing scenario in csharp editing speedometer test ***
*** Memory allocations under RegularCompilationTracker.WithDoNotCreateCreationPolicy ***
