Update benchmark.yml by lofcz · Pull Request #31 · lofcz/FastCloner

lofcz · 2026-03-04T17:29:13Z

Note

Medium Risk
Touches core cloning hot paths (type-dispatch caching and class-array cloning fast paths), which could affect clone correctness/perf for polymorphic arrays and override scenarios. CI/reporting changes are low risk but could alter baseline selection and diff presentation.

Overview
Updates the benchmark GitHub Actions workflow to run only ubuntu-latest by default (while keeping optional multi-OS runs for workflow_dispatch) and switches baseline retrieval to gh run download using the originating run id.

Improves benchmark diff output formatting by moving Status to the first column and rendering statuses as symbols (🔴/🟢/🟡/🆕/⚪), and slightly simplifies CSV time parsing by consolidating microsecond unit handling.

Optimizes cloning internals by expanding TypeCloneDispatchCache2 from a 2-entry to a 4-entry promotion cache and adding a fast path in Clone1DimArrayClassInternal to directly copy safe/primitive/enum elements when no optional type behavior overrides are active.

^{Written by Cursor Bugbot for commit f9a0102. This will update automatically on new commits. Configure here.}

Copilot

Pull request overview

Updates the benchmark GitHub Actions workflow to avoid scheduling unnecessary OS matrix jobs except when manually dispatched, reducing CI load for push/PR events.

Changes:

Make strategy.matrix.os dynamic so push/PR runs only execute on ubuntu-latest, while workflow_dispatch can target all OSes.
Simplify the run-gating PowerShell logic by defaulting $shouldRun to true and only applying input-based gating for manual dispatches.

Comments suppressed due to low confidence (1)

.github/workflows/benchmark.yml:53

$shouldRun now defaults to $true. If a new/typo OS value ever gets added to the matrix without updating the switch, it will run unintentionally on workflow_dispatch. Safer pattern is to default to $false within the dispatch branch (or add a default case) so only explicitly handled OS values are allowed to proceed.

        $shouldRun = $true

        if ($eventName -eq "workflow_dispatch") {
          switch ($os) {
            "ubuntu-latest" { $shouldRun = "${{ inputs.run_ubuntu }}" -eq "true"; break }
            "windows-latest" { $shouldRun = "${{ inputs.run_windows }}" -eq "true"; break }
            "macos-latest" { $shouldRun = "${{ inputs.run_macos }}" -eq "true"; break }

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-03-04T17:31:39Z

      fail-fast: false
      matrix:
-        os: [ubuntu-latest, windows-latest, macos-latest]
+        os: ${{ fromJson(github.event_name == 'workflow_dispatch' && '["ubuntu-latest","windows-latest","macos-latest"]' || '["ubuntu-latest"]') }}


For workflow_dispatch, the matrix always schedules all 3 OS jobs even when an input (e.g. run_windows) is false; those jobs still provision a runner just to run the gate step and then skip the rest. Consider building the matrix.os list from the dispatch inputs so disabled OS values are omitted entirely, reducing runner usage and queue time.

Suggested change

os: ${{ fromJson(github.event_name == 'workflow_dispatch' && '["ubuntu-latest","windows-latest","macos-latest"]' || '["ubuntu-latest"]') }}

os: ${{ fromJson(github.event_name == 'workflow_dispatch' && format('[{0}{1}{2}]',

inputs.run_ubuntu && '"ubuntu-latest"' || '',

inputs.run_windows && (inputs.run_ubuntu && ',"windows-latest"' || '"windows-latest"') || '',

inputs.run_macos && ((inputs.run_ubuntu || inputs.run_windows) && ',"macos-latest"' || '"macos-latest"') || ''

) || '["ubuntu-latest"]') }}

cursor

Cursor Bugbot has reviewed your changes and found 1 potential issue.

Bugbot Free Tier Details

You are on the Bugbot Free tier. On this plan, Bugbot will review limited PRs each billing cycle.

To receive Bugbot reviews on all of your PRs, visit the Cursor dashboard to activate Pro and start your 14-day free trial.

Bugbot Autofix prepared a fix for the issue found in the latest run.

✅ Fixed: Safe-type shortcut misses boxed value type exclusion
- Added !runtimeType.IsValueType check to the fast-path condition in Clone1DimArrayClassInternal to exclude boxed value types, matching the pattern used in CloneClassShallowAndTrack and BuildTypeMetadata.

…ClassInternal The fast-path optimization at line 979 only checked CanReturnSameObject(runtimeType) without excluding value types. When T is object or an interface, array elements can be boxed structs that must be deep-cloned (boxed structs can be mutated). This fix adds the same !runtimeType.IsValueType check used in CloneClassShallowAndTrack and BuildTypeMetadata to ensure boxed value types are properly deep-cloned.

github-actions · 2026-03-04T18:03:00Z

Deep Clone Benchmarks

OS: ubuntu-latest
Generated (UTC): 2026-03-04 18:35:33

Current FastCloner vs DeepCloner

Benchmark	DeepCloner	FastCloner	Delta Time	DC Alloc	FC Alloc	Delta Alloc
SmallObject	82.44 ns	74.28 ns	-10% faster	184 B	48 B	-74% less
FileSpec	494.09 ns	298.40 ns	-40% faster	920 B	416 B	-55% less
StringArray_1000	517.79 ns	518.35 ns	~same	8,160 B	8,024 B	~same
SmallObjectWithCollections	646.36 ns	361.67 ns	-44% faster	1,096 B	576 B	-47% less
DynamicWithDictionary	1,354.53 ns	1,074.06 ns	-21% faster	2,712 B	1,600 B	-41% less
MediumNestedObject	1,693.81 ns	1,110.34 ns	-34% faster	3,416 B	1,616 B	-53% less
DynamicWithNestedObject	1,887.28 ns	1,313.28 ns	-30% faster	3,560 B	1,776 B	-50% less
DynamicWithArray	5,794.83 ns	5,023.64 ns	-13% faster	8,800 B	2,744 B	-69% less
LargeEventDocument_10MB	69,019.25 ns	38,058.96 ns	-45% faster	129,792 B	49,824 B	-62% less
ObjectList_100	171,149.39 ns	107,334.11 ns	-37% faster	318,888 B	149,816 B	-53% less
ObjectDictionary_50	841,546.87 ns	183,984.76 ns	-78% faster	549,660 B	218,768 B	-60% less
LargeLogBatch_10MB	6,162,714.03 ns	3,671,896.53 ns	-40% faster	3,564,661 B	2,649,104 B	-26% less

FastCloner vs latest `next` baseline

Baseline generated (UTC): 2026-03-04 17:24:17
Regression thresholds: time > 5%, alloc > 5%

Status	Benchmark	FC Time (Baseline)	FC Time (Current)	Delta Time	FC Alloc (Baseline)	FC Alloc (Current)	Delta Alloc
🟢	DynamicWithArray	6,359.98 ns	5,023.64 ns	-21% faster	2,840 B	2,744 B	-3% less
⚪	DynamicWithDictionary	1,024.11 ns	1,074.06 ns	+5% slower	1,600 B	1,600 B	~same
⚪	DynamicWithNestedObject	1,344.40 ns	1,313.28 ns	-2% faster	1,776 B	1,776 B	~same
⚪	FileSpec	297.28 ns	298.40 ns	~same	416 B	416 B	~same
⚪	LargeEventDocument_10MB	37,918.57 ns	38,058.96 ns	~same	49,824 B	49,824 B	~same
🔴	LargeLogBatch_10MB	3,118,356.96 ns	3,671,896.53 ns	+18% slower	2,649,094 B	2,649,104 B	~same
⚪	MediumNestedObject	1,121.19 ns	1,110.34 ns	~same	1,616 B	1,616 B	~same
⚪	ObjectDictionary_50	178,335.37 ns	183,984.76 ns	+3% slower	218,768 B	218,768 B	~same
⚪	ObjectList_100	107,796.83 ns	107,334.11 ns	~same	149,816 B	149,816 B	~same
⚪	SmallObject	73.12 ns	74.28 ns	~same	48 B	48 B	~same
🟢	SmallObjectWithCollections	386.55 ns	361.67 ns	-6% faster	576 B	576 B	~same
🟢	StringArray_1000	623.53 ns	518.35 ns	-17% faster	8,024 B	8,024 B	~same

Regressions

LargeLogBatch_10MB: time +18% slower, alloc ~same

Improvements

DynamicWithArray: time -21% faster, alloc -3% less
SmallObjectWithCollections: time -6% faster, alloc ~same
StringArray_1000: time -17% faster, alloc ~same

Mixed changes

none

…into chore-ci-bench

Copilot

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-03-04T18:15:49Z

    private struct TypeCloneDispatchCache2
    {
        private Type? typeA;
        private FastClonerCache.TypeCloneMetadata? metadataA;
        private Func<object, FastCloneState, object>? recursiveA;
        private Func<object, FastCloneState, object>? worklistA;
        private Type? typeB;
        private FastClonerCache.TypeCloneMetadata? metadataB;
        private Func<object, FastCloneState, object>? recursiveB;
        private Func<object, FastCloneState, object>? worklistB;
+        private Type? typeC;
+        private FastClonerCache.TypeCloneMetadata? metadataC;
+        private Func<object, FastCloneState, object>? recursiveC;
+        private Func<object, FastCloneState, object>? worklistC;
+        private Type? typeD;
+        private FastClonerCache.TypeCloneMetadata? metadataD;
+        private Func<object, FastCloneState, object>? recursiveD;
+        private Func<object, FastCloneState, object>? worklistD;


TypeCloneDispatchCache2 now holds 4 cached entries (A–D). The "2" suffix is misleading and makes the cache behavior harder to reason about when revisiting this code. Consider renaming the struct to reflect its actual capacity (e.g., TypeCloneDispatchCache4) or removing the numeric suffix if it no longer conveys meaning.

Update benchmark.yml

2eceb35

Copilot AI review requested due to automatic review settings March 4, 2026 17:29

Copilot started reviewing on behalf of lofcz March 4, 2026 17:29 View session

Copilot AI reviewed Mar 4, 2026

View reviewed changes

fix

05a0657

cursor Bot reviewed Mar 4, 2026

View reviewed changes

Comment thread src/FastCloner/Code/FastClonerGenerator.cs

Copilot AI review requested due to automatic review settings March 4, 2026 17:59

lofcz removed the request for review from Copilot March 4, 2026 17:59

ups

90e4af7

Copilot AI review requested due to automatic review settings March 4, 2026 18:11

Merge branch 'chore-ci-bench' of https://github.com/lofcz/FastCloner …

4f4cb4d

…into chore-ci-bench

Copilot started reviewing on behalf of lofcz March 4, 2026 18:12 View session

Copilot AI reviewed Mar 4, 2026

View reviewed changes

Update BenchmarkDiffReporter.cs

f9a0102

lofcz mentioned this pull request Mar 4, 2026

FastCloner as DeepCloner replacement FoundatioFx/Foundatio#469

Merged

6 tasks done

lofcz merged commit cc0e4c0 into next Mar 4, 2026
12 checks passed

Navigation Menu

Repository navigation

Update benchmark.yml - #31

Update benchmark.yml#31
lofcz merged 6 commits into
nextfrom
chore-ci-bench

lofcz commented Mar 4, 2026 •

edited by cursor Bot

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Mar 4, 2026

Uh oh!

cursor Bot left a comment •

edited

Loading

Uh oh!

Uh oh!

github-actions Bot commented Mar 4, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Mar 4, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

-        os: ${{ fromJson(github.event_name == 'workflow_dispatch' && '["ubuntu-latest","windows-latest","macos-latest"]' || '["ubuntu-latest"]') }}
+        os: ${{ fromJson(github.event_name == 'workflow_dispatch' && format('[{0}{1}{2}]',
+          inputs.run_ubuntu && '"ubuntu-latest"' || '',
+          inputs.run_windows && (inputs.run_ubuntu && ',"windows-latest"' || '"windows-latest"') || '',
+          inputs.run_macos && ((inputs.run_ubuntu || inputs.run_windows) && ',"macos-latest"' || '"macos-latest"') || ''
+        ) || '["ubuntu-latest"]') }}

Conversation

lofcz commented Mar 4, 2026 • edited by cursor Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Mar 4, 2026

Choose a reason for hiding this comment

Uh oh!

cursor Bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

github-actions Bot commented Mar 4, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Deep Clone Benchmarks

Current FastCloner vs DeepCloner

FastCloner vs latest next baseline

Regressions

Improvements

Mixed changes

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Mar 4, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

lofcz commented Mar 4, 2026 •

edited by cursor Bot

Loading

cursor Bot left a comment •

edited

Loading

github-actions Bot commented Mar 4, 2026 •

edited

Loading

FastCloner vs latest `next` baseline