High Contention in MemoryCache _cacheSize Updates Causing Performance Degradation

Open
#111,959 5 comments 5 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Assessment

Difficulty
5/5
Estimated time
Over a week
Newbie friendliness
35/100
Issue type
Bug
Clarity
Mostly clear
Activity status
Stale
Tech stack
csharp

Research direction

Start by locating MemoryCache and the UpdateCacheSizeExceedsCapacity method, then review how _cacheSize is updated during concurrent cache writes. Reproduce the provided Parallel.For example on .NET 9 and measure CPU usage and throughput under contention. Done means a validated approach reduces contention without changing cache capacity behavior.

Written by the indexing model from the issue text.

Description

area-Extensions-Caching help wanted tenet-performance
Description

The current implementation of MemoryCache updates _cacheSize using Interlocked.CompareExchange in a retry loop (up to 100 times). This creates high contention when multiple threads attempt to modify _cacheSize, leading to increased CPU usage and reduced throughput under high concurrency.

using System;
using System.Threading;
using System.Threading.Tasks;
using Microsoft.Extensions.Caching.Memory;
using Microsoft.Extensions.Options;

var cache = new MemoryCache(Options.Create(new MemoryCacheOptions { SizeLimit = 1000 }));

Parallel.For(0, 100, i =>
{
    cache.Set(i, i, new MemoryCacheEntryOptions { Size = 10 });
});

Console.WriteLine($"Cache Count: {cache.Count}"); 

Issue: When running with a high number of concurrent threads, CPU usage spikes and cache performance degrades due to frequent retries in UpdateCacheSizeExceedsCapacity.

Configuration

.NET Version: .NET 9
OS: Windows 11
Architecture: x64
Machine Specs: 16-core CPU, 16GB RAM

Regression?

Not a regression but a long-standing inefficiency in MemoryCache.

Data
Analysis

Problem:

  • Interlocked.CompareExchange retried up to 100 times under contention.

  • Causes CPU spikes when multiple threads modify _cacheSize.

  • Locks too frequently, slowing down high-throughput applications.

Proposed Fix:
✅ Use ConcurrentQueue<long> for batch updates instead of per-entry updates.
✅ Process updates asynchronously in a background task, reducing lock contention.
✅ Minimize atomic operations on _cacheSize, improving cache scalability.

Dominant language
C#
Stars
18.3k
Forks
5.6k
PR merge metrics
PR metrics pending

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from dotnet/runtime

All issues in dotnet/runtime

Similar issues

More C# issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.