About C# LINQ Performance

Table of Contents

• Advantages of LINQ
• Usage Rules
• Benchmark - Loops
• Benchmark - Filtering
• Benchmark - Transformation
• How Much GC Does LINQ Create?
• Conclusion

---

Advantages of LINQ

• LINQ (Language Integrated Query) is a C# feature set for writing concise and convenient code.
  Developers can use query syntax to perform filtering, sorting, and grouping on data sources
  with minimal code.

Notes When Using LINQ

• Most LINQ operators create intermediate buffers (a kind of array), and all of that becomes garbage. 👉 GC occurs
• Even when one or two LINQ operators are enough, splitting logic into many LINQ calls for readability can degrade performance.
• Operators such as Average, Count, OrderBy, and Reverse iterate the source sequence, so they are less efficient from a performance perspective.

So before using LINQ, think through these rules.

Usage Rules

1. Is this a situation where rapid implementation matters (prototype, TDD)? - Use LINQ
2. Is this code not performance-sensitive? - Use LINQ
3. Is LINQ’s convenience/readability truly needed here? - Use LINQ
4. Is this called every frame or performance-critical? - Do not use LINQ

LINQ Performance Benchmark

• Benchmark library: BenchMark
• Test code: TestCode
• Original article: LinqPerformance

Benchmark - Loops

Return the number of Customers whose age is below 18

For / Foreach

// Performance: fast (no GC)
int count = 0;
foreach (var c in customers)
{
    if (c.Age < 18)
        count++;
}
return count;

LINQ Count

// Performance: slower (GC occurs)
// Count() internally
// iterates the entire sequence
return customers
    .Count(c => c.Age < 18);

1. For / Foreach

2. LINQ Count / Predicate Count
   

Compared to for/foreach, LINQ has some overhead (indirect extra time, memory, and resources for functionality).
Filtering then counting is relatively close to for, but plain Count that returns sequence length shows notable performance degradation.

Benchmark - Filtering

Return a list that is a subset of customers with age 18 or higher

1. For / Foreach

2. LINQ

It shows very similar performance to for/foreach implementation. About 4ms slower than for/foreach.

Example code actually used when filtering RaycastAll results and checking tags

For filtering scenarios, LINQ made the code much more concise and readable than typical for/foreach.

Benchmark - Transformation

Return a list of customers where age is represented in months instead of years

1. For / Foreach

2. LINQ

How Much GC Does LINQ Create?

See the link below for test conditions.

• Original article: Just How Much Garbage Does LINQ Create?

Result

1. GC Alloc Count
   
   

Microsoft Docs: Performance Recommendations for Unity

Performance recommendations for Unity

LINQ vs For - Performance Comparison Chart

Relative execution time (lower is faster, for = 1.0 baseline)

Conclusion

1. LINQ Count is significantly slower.
2. Filtering (Where) and transformation (Select) are similar to for in performance.
3. LINQ allows very concise one- or two-line code.
4. Debugging is harder.
5. GC allocation occurs; overusing LINQ can become hard to manage.
6. Avoid LINQ in per-frame calls or other performance-sensitive paths.