HackerReposA FIELD GUIDE TO OPEN SOURCE
ProSponsor the shelf♧ 0
← Back to top all-time
ben-manes
caffeine
Java2015–2025
ben-manes / caffeine

caffeine

A high performance caching library for Java

19Distinct people
51Mentions
1Explicit recommenders
17.9KGitHub stars
REPOSITORY AT A GLANCE

What to know before exploring

What it is

A high performance caching library for Java

GitHub

Where it is used

No documented product use was verified in the last source check.

What people are saying

This is sometimes referred to as the "admission policy" of a cache and can have a large effect at increasing the hit ratio. Most policies try to minimize the impact of low…

NovaX used a recommendation phrase. One indexed comment matches. This is a phrase match, not a verdict.

No classified warning is attached to this page. Read the discussions before treating silence as approval.

Also in these discussions

No other repository is repeatedly linked in the same discussions.

This week on Hacker News

Nov 9–Nov 16 UTC · Direct repository links and separately labeled related project stories. Indexed through 11/16/2026, 4:59:52 AM UTC.

No direct repository links were found in the indexed HN discussions over the past seven days. Older discussions are available below.

Mentions over time

HN mentions by year51 dated mentions

Observed mentions: 2015-02-24 07:16:52 UTC to 2025-03-06 01:57:40 UTC.

024682015-01-01 00:00:00 UTC to 2016-01-01 00:00:00 UTC (exclusive): 7 mentions; 3 submissions, 4 comments20152016-01-01 00:00:00 UTC to 2017-01-01 00:00:00 UTC (exclusive): 5 mentions; 1 submissions, 4 comments2017-01-01 00:00:00 UTC to 2018-01-01 00:00:00 UTC (exclusive): 2 mentions; 0 submissions, 2 comments2018-01-01 00:00:00 UTC to 2019-01-01 00:00:00 UTC (exclusive): 3 mentions; 0 submissions, 3 comments2019-01-01 00:00:00 UTC to 2020-01-01 00:00:00 UTC (exclusive): 8 mentions; 0 submissions, 8 comments2020-01-01 00:00:00 UTC to 2021-01-01 00:00:00 UTC (exclusive): 1 mentions; 0 submissions, 1 comments20202021-01-01 00:00:00 UTC to 2022-01-01 00:00:00 UTC (exclusive): 7 mentions; 0 submissions, 7 comments2022-01-01 00:00:00 UTC to 2023-01-01 00:00:00 UTC (exclusive): 2 mentions; 0 submissions, 2 comments2023-01-01 00:00:00 UTC to 2024-01-01 00:00:00 UTC (exclusive): 7 mentions; 0 submissions, 7 comments2024-01-01 00:00:00 UTC to 2025-01-01 00:00:00 UTC (exclusive): 3 mentions; 0 submissions, 3 comments2025-01-01 00:00:00 UTC to 2026-01-01 00:00:00 UTC (exclusive): 6 mentions; 0 submissions, 6 comments2025

Counts come from the repository’s dated source evidence. Index cutoff: 2026-11-02 08:16:59 UTC.

View mention counts and date ranges
UTC ranges include their start and exclude their end. Select a period to filter the source evidence.
PeriodStart (UTC)End (exclusive, UTC)MentionsSubmissionsComments
20152015-01-01 00:00:00 UTC2016-01-01 00:00:00 UTC734
20162016-01-01 00:00:00 UTC2017-01-01 00:00:00 UTC514
20172017-01-01 00:00:00 UTC2018-01-01 00:00:00 UTC202
20182018-01-01 00:00:00 UTC2019-01-01 00:00:00 UTC303
20192019-01-01 00:00:00 UTC2020-01-01 00:00:00 UTC808
20202020-01-01 00:00:00 UTC2021-01-01 00:00:00 UTC101
20212021-01-01 00:00:00 UTC2022-01-01 00:00:00 UTC707
20222022-01-01 00:00:00 UTC2023-01-01 00:00:00 UTC202
20232023-01-01 00:00:00 UTC2024-01-01 00:00:00 UTC707
20242024-01-01 00:00:00 UTC2025-01-01 00:00:00 UTC303
20252025-01-01 00:00:00 UTC2026-01-01 00:00:00 UTC606

What Hacker News says

6 sources
This was a good attempt but flawed. 1. You should use JMH to handle warmup, jit, timings, averaging of runs, etc. You might enjoy the following talks on the subject, though I am sure there are other gems out there. - Performance Anxiety: https://wiki.jvmlangsummit.com/images/1/1d/PerformanceAnxiet... - Benchmarking for Good: https://www.youtube.com/watch?v=SKPdqgD1I2U - (The Art of) (Java) Benchmarking: https://shipilev.net/talks/j1-Oct2011-21682-benchmarking.pdf - A Crash Course in Modern Hardware: https://www.youtube.com/watch?v=OFgxAFdxYAQ - How NOT to Write a Microbenchmark https://www.slideshare.net/slideshow/2002-microbenchmarks/28... - Anatomy of a flawed microbenchmark https://web.archive.org/web/20110513090823/http://www.ibm.co... 2. An uncontended lock is basically free, as you noticed. 3. Guava does implement the Map interface
Read the rest
via asMap(). It also have a concurrencyLevel to adjust its performance. It is based on Java 5's CHM. 4. Caffeine is a Guava Cache rewrite based on Java 8's CHM rewrite, plus many learnings. You might look at its benchmarks for ideas: https://github.com/ben-manes/caffeine/wiki/Benchmarks 5. Data structures are surprisingly tricky. For example see this analysis showing an accidental misunderstanding degrading an LRU to O(n) eviction. https://gist.github.com/ben-manes/6312727adfa2235cb7c5e25cae... 6. It is important to remember that the goal of a benchmark is never which is faster or by how much. It is to ask (a) is it fast enough? (b) might I reach a point where it will not be? and (c) does it degrade unexpectedly? == This is to say the winner is of little interest, once all the choices are good enough then it is about usability, features, documentation, friendliness, and so on.
In Microbenchmarks: HashMap, ConcurrentHashMap, and Guava Caches
You might be interested in this thread [1] where I described an idea for how to incorporate the latency penalty into the eviction decision. A developer even hacked a prototype that showed promise. The problem is that there is not enough variety in the available trace data to be confident that a design isn't overly fit to a particular workload and doesn't generalize. As more data sets become available it will become possible to experiment with ideas and fix unexpected issues until a correct, simple, elegant design emerges. [1] https://github.com/ben-manes/caffeine/discussions/1744
In Analyzing the codebase of Caffeine, a high performance caching library
As the article mentions, Caffeine's approach is to monitor the workload and adapt to these phase changes. This stress test [1] demonstrates shifting back and forth between LRU and MRU request patterns, and the cache reconfiguring itself to maximize the hit rate. Unfortunately most policies are not adaptive or do it poorly. Thankfully most workloads are a relatively consistent pattern, so it is an atypical worry. The algorithm designers usually have a target scenario, like cdn or database, so they generally skip reporting the low performing workloads. That may work for a research paper, but when providing a library we cannot know what our users workloads are nor should we expect engineers to invest in selecting the optimal algorithm. Caffeine's adaptivity removes this burden and broaden its applicability, and other language ecosystems have
Read the rest
been slowly adopting similar ideas in their caching libraries. [1] https://github.com/ben-manes/caffeine/wiki/Efficiency#adapti...
In Analyzing the codebase of Caffeine, a high performance caching library
I tried to reimplement Linux’s algorithm in [1], but I cannot be sure about correctness. They adjust the fixed sizes at construction based on device’s total memory, so it varies if a phone or server. This fast trace simulation in the CI [2] may be informative (see DClock). Segmentation is very common, where algorithms differ by how they promote and how/if they adapt the sizes. [1] https://github.com/ben-manes/caffeine/blob/master/simulator/... [2] https://github.com/ben-manes/caffeine/actions/runs/130865965...
In Analyzing the codebase of Caffeine, a high performance caching library

MAKE YOURSELF AT HOME

A few shortcuts

/
Search the collection
?
Open this guide
Esc
Close dialogs and navigation

Your saved repositories and interests stay in this browser.