Top all-time · #1808
geekbench.com
Company blog · geekbench.com/bloghardwareprogramming
geekbench.com is the company blog at geekbench.com, writing about hardware and programming. Since 2017, 41 different Hacker News readers have linked it 50 times, most often “iPhone Performance Degrades as Battery Ages”. No feed was found, so read it on the site itself.
Rarely recommended · 1 of 33 sampled comments recommend it?
Why HN links it
HN readers link Geekbench's blog posts and workload documents to explain what the benchmark tests and how scores are produced. Posts on iPhone throttling and CPU comparisons are cited as evidence in discussions.
- Recommended for · 1 comments
- One commenter recommends reading the Geekbench 6 internals document, saying it explains the rationale for the shift to the shared-task multi-threading model.1
Summarised by Claude from a sample of 33 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Most-linked posts?
- /doc/geekbench5-cpu-workloads.pdf10 people · 10 mentions · first linked 2019-09-21
- iPhone Performance Degrades as Battery Ages8 people · 8 mentions · first linked 2017-12-18
- Geekbench 6 Benchmark Internals [pdf]6 people · 8 mentions · first linked 2023-10-04
- /doc/geekbench4-cpu-workloads.pdf4 people · 4 mentions · first linked 2018-06-14
- MacBook Pro (Mid 2018) Throttling may be due to GPU2 people · 2 mentions · first linked 2018-07-20
- Intel's Binary Optimization Tool may be cheating on Geekbench 62 people · 2 mentions · first linked 2026-03-25
- Analyzing Geekbench 6 under Intel's BOT2 people · 2 mentions · first linked 2026-03-31
What Hacker News says?
One thing that Apple hasn't addressed is if the Geekbench results [1] over-represents or exaggerates the the throttling effect. Since Geekbench pushes the iPhone to its limits for the duration of the benchmark, iOS needs to throttle performance to prevent shutdown while being tested. This results in lower-than-expected test scores every time. But how much does this map to real-life, visible performance drop? Edit: A…
Comment linking /blog/2017/12/iphone-performance-and-battery-age/
on “A Message to Our Customers about iPhone Batteries and Performance”Submitted iPhone Performance Degrades as Battery Ages as a story · 456 points
HN discussionLooks like it has (as of 6 hours ago) https://www.geekbench.com/blog/2020/11/geekbench-53/
Comment linking /blog/2020/11/geekbench-53/
on “Apple Silicon M1 chip in MacBook Air outperforms high-end 16-inch MacBook Pro”Submitted MacBook Pro (Mid 2018) Throttling may be due to GPU as a story · 28 points
HN discussion>my PPC Mac mini was more performant than the first Intel mini by far That's true only if "Apple using CPU I deem cool" counts as performance. Otherwise, benchmarks say that first Intel Mac Mini was almost twice as fast than the latest Mac Mini with PowerPC. https://www.geekbench.com/blog/2009/01/mac-performance-janua...
Comment linking /blog/2009/01/mac-performance-january-2009/
on “Apple aims to sell Macs with its own chips starting in 2021”Submitted Geekbench AI 1.0 as a story · 26 points
HN discussionFWIW You can review the methodology that they use here: https://www.geekbench.com/doc/geekbench4-cpu-workloads.pdf They don't provide source though. That's a detriment. It would be nice if their benchmarks were publicly audit-able.
Comment linking /doc/geekbench4-cpu-workloads.pdf
on “Not Only Is the Mac Mini Outdated, It's No Longer Mini”Submitted Geekbench 7 as a story · 13 points
HN discussionIn Geekbench5, they benchmark html5, SQLite reads, pdf rendering, text rendering etc. Seems somewhat relevant. They are providing an upper bound on this activity, and that’s good to know. https://www.geekbench.com/doc/geekbench5-cpu-workloads.pdf
Comment linking /doc/geekbench5-cpu-workloads.pdf
on “Apple Silicon M1 Emulating x86 Is Still Faster Than Every Other Mac”Submitted Geekbench: iPhone Performance and Battery Age as a story · 10 points
HN discussionRight, that's also my guess. What I meant was that their document describing the v4 tests[1] mentions SSE and AVX, while the v5 version[2] does not. Does it mean that the v5 tests don't cover them? I found find it surprising for a benchmarking tool but they really were mentioned before and now they're not, hence my question. If indeed v5 tests are not particularly about specialized performance but more about real-wo…
Comment linking /doc/geekbench5-cpu-workloads.pdf
on “Apple Silicon M1 Emulating x86 Is Still Faster Than Every Other Mac”Submitted iPhone Performance and Battery Age as a story · 8 points
HN discussionGeekbench tries to simulate real world situations. HTML and PDF rendering, SQLite, etc https://www.geekbench.com/doc/geekbench5-cpu-workloads.pdf
Comment linking /doc/geekbench5-cpu-workloads.pdf
on “Apple Silicon M1 Emulating x86 Is Still Faster Than Every Other Mac”Submitted Analyzing Geekbench 6 under Intel's BOT as a story · 3 points
HN discussionCalling Geekbench a "single benchmark" is misleading because it runs 25 different types of workloads. So AMD doing well here means it does well compared to Intel in 25 different areas when they are all averaged: • Cryptography Workloads - AES • Integer Workloads - LZMA Compression - JPEG Compression - Canny - Lua - Dijkstra - SQLite - HTML 5 Parse - HTML 5 DOM - Histogram Equalization - PDF Rendering - LLVM - Camera…
Comment linking /doc/geekbench4-cpu-workloads.pdf
on “$499 AMD Ryzen 9 3900X Almost as Fast as $2000 Intel Core I9-9980XE”Submitted Geekbench 6.4 released with RISC-V Vector (RVV) and improved ARM SVE support as a story · 3 points
HN discussionYou can look at the Geekbench 6 component. https://www.geekbench.com/doc/geekbench6-benchmark-internals... (page 18)
Comment linking /doc/geekbench6-benchmark-internals.pdf
on “AMD Unveils Ryzen 9000 CPUs for Desktop, Zen 5”Submitted Geekbench 5 as a story · 2 points
HN discussion> Clang is a compiler front end for the programming languages C, C++, Objective-C, Objective-C++, OpenMP, OpenCL, and CUDA. It uses LLVM as its back end. The Clang workload compiles a 1,094 line C source file (of which 729 lines are code). The workload uses AArch64 as the target architecture for code generation. 729 LOC within a single file doesn't represent ANYTHING about real world compilation. It neatly fits with…
Comment linking /doc/geekbench5-cpu-workloads.pdf
on “ARM Mac Impact on Intel”Submitted Analyzing Geekbench 6 Under Intel's Binary Optimization Tool (Bot) as a story · 1 points
HN discussionFWIW you've linked Geekbench 4; Geekbench 5 is https://www.geekbench.com/doc/geekbench5-cpu-workloads.pdf I agree these benchmarks have issues, but there's a consistent trend (eg. AnandTech's SPEC results are similar—they use active cooling to avoid the question about thermals) and the most reliable benchmarks, like compilers, aren't outliers.
Comment linking /doc/geekbench5-cpu-workloads.pdf
on “ARM Mac Impact on Intel”Submitted Intel's Binary Optimization Tool may be cheating on Geekbench 6 as a story · 1 points
HN discussiongeekbench is a horrid benchmark.The amount of info on how they run the tests is extremely limited (mostly what lib they opted to use). The tests are very short, no power or thermally limited. I'd try and link to their methodology[0]: for instance: Integer WorkloadsLZMA CompressionLZMA (Lempel-Ziv-Markov chain algorithm) is a lossless compression algorithm. The algorithm uses a dictionary compression scheme (the dict…
Comment linking /doc/geekbench4-cpu-workloads.pdf
on “ARM Mac Impact on Intel”Submitted The secret sauce of Geekbench CPU benchmark [pdf] as a story · 1 points
HN discussionThey don't benchmark apps, but test on a set of standard workloads and algorithms. https://www.geekbench.com/doc/geekbench5-cpu-workloads.pdf
Comment linking /doc/geekbench5-cpu-workloads.pdf
on “AMD’s Mobile Revival: Redefining the Notebook Business with the Ryzen 9 4900HS”Submitted Geekbench 6 Benchmark Internals [pdf] as a story · 1 points
HN discussionCrypto stuff never uses floating point. It might have built-in instructions like AES-NI, though. In the test geekbench consider that - integers ones[0] and definitely not floating point. Media encoding tend not to use floating point, either. (decoders tend to be built-in in the GPUs with designated hardware, so testing on CPU is quite pointless) [0]: https://www.geekbench.com/doc/geekbench5-cpu-workloads.pdf
Comment linking /doc/geekbench5-cpu-workloads.pdf
on “The 2020 iPhone SE”Submitted Geekbench 5: Compute Workloads [pdf] as a story · 1 points
HN discussionhttps://www.geekbench.com/doc/geekbench5-cpu-workloads.pdf Looks like a variety of workloads, but lots of things that can be greatly improved with co-procesors. Like cryptography, image compression, image processing, etc. "Geekbench 5 scores are calibrated against a baseline score of 1,000 (which is the score of a Dell Precision 3430 with a Core i3-8100 processor)"
Comment linking /doc/geekbench5-cpu-workloads.pdf
on “The 2020 iPhone SE”This post gives us an estimate for what percentage of the iPhones are getting throttled by this issue. I think it's reasonable to use this as a proxy to estimate how many units were at risk of shutting down and it is the minority of units. Happily it appears both of your devices are in the majority of units which don't exhibit this issue. https://www.geekbench.com/blog/2017/12/iphone-performance-an...
Comment linking /blog/2017/12/iphone-performance-and-battery-age/
on “Apple Says It Slows Older iPhones to Save Their Battery Life”
