CompactifAI Inference API

2 points by Compactifai 1h ago 0 comments

Ask HN: The government of my country blocked VPN access. What should I use?

59 points by rickybule 1h ago 54 comments

Ask HN: Why hasn't x86 caught up with Apple M series?

430 points by stephenheron 2d ago 613 comments

Ask HN: Anyone working on bringing software back from US clouds?

4 points by sam_lowry_ 4h ago 4 comments

Ask HN: Windows 11 Update Fail – Linux Distro Suggestions?

3 points by giardini 1h ago 2 comments

Petition to stop Google from restricting sideloading and FOSS apps

252 points by nativeforks 7h ago 143 comments

Ask HN: How can I recover and run my old mobile game from the 2010s?

56 points by diasks2 3d ago 48 comments

Anthropick.com Redirects to ChatGPT

12 points by nerdcortex 14h ago 1 comments

Ask HN: GitHub Copilot down?

58 points by thecopy 21h ago 20 comments

Ask HN: Did modern AI's coding abilities make you lose interest in programming?

29 points by amichail 1d ago 41 comments

Ask HN: What to Do with Old iPads?

30 points by owenmakes 2d ago 19 comments

Ask HN: What to Learn for Math for Modeling?

2 points by shivajikobardan 1d ago 0 comments

Ask HN: How to Learn to Build Agentic AI Systems (Like Claude Code)

6 points by hhimanshu 20h ago 3 comments

Ask HN: What are the best Google alternatives in 2025?

13 points by eel 20h ago 4 comments

Units of Economics of LLMs. Reply to Ed Zitron's "AI Is a Money Trap"

6 points by tudorizer 1d ago 6 comments

Ask HN: What measures are you taking to stop AI crawlers?

6 points by kjok 22h ago 14 comments

Ask HN: Does sentience put stress on the brain?

7 points by trinsic2 1d ago 4 comments

Ask HN: Is backlink trading still a problem worth solving?

3 points by sathishn 23h ago 1 comments

Ask HN: Is there a temp phone number like temp email?

9 points by piratesAndSons 1d ago 11 comments

Ask HN: Best codebases to study to learn software design?

103 points by pixelworm 4d ago 90 comments

Out of curiosity: what kind of people use this "forum" (I mean Hacker News)?

21 points by adinhitlore 1d ago 27 comments

Ask HN: Why are so many services rejecting Google Voice numbers for signups?

12 points by electric_muse 1d ago 19 comments

Ask HN: What should I use to run React Native tests on a device?

4 points by dboreham 1d ago 0 comments

Tell HN: any reasonably used DB will likely outlast the programs using it

3 points by dvrp 12h ago 1 comments

Stop squashing your commits. You're squashing your AI too

4 points by jannesblobel 2d ago 10 comments

Ask HN: Are AI filters becoming stricter than society itself?

29 points by tsevis 4d ago 16 comments

Ask HN: Any experienced devs who use AI extensively in their work?

3 points by atleastoptimal 1d ago 4 comments

Ask HN: I just abandoned my PyCharm subscription, what should I use now?

22 points by jmward01 3d ago 23 comments

Ask HN: How are you attributing your AI usage when developing software?

4 points by hartleybrody 1d ago 4 comments

Patient Lisp Hacker Seeks Same for Long Walks Through IPL-V Code

7 points by abrax3141 2d ago 2 comments

Ask HN: Has anyone else used online communities that are archetypically "savvy"?

2 points by Use 2d ago 6 comments

High rate of LLM (GPT5) hallucinations in dense stats domains (cricket)

3 points by sp1982 2d ago 2 comments

Ask HN: What are you working on (August 2025)?

10 points by iryndin 2d ago 15 comments

Ask HN: How do you find early stage startups to join

44 points by gavino 6d ago 23 comments

429 Too Many Requests from registry.npmjs.org

2 points by mirekrusin 2d ago 2 comments

Why is AI search still bad at trust and context?

2 points by zyruh 1d ago 1 comments

Ask HN: Best Marketplaces for Used Servers?

8 points by bloudermilk 3d ago 11 comments

Ask HN: Does using public transportation make you more creative than driving?

7 points by amichail 3d ago 11 comments

How can a mutex in Wine be faster than a native one on Linux

3 points by lh_mouse 2d ago 2 comments

Ask HN: What is wrong with modern software development

6 points by fzeindl 2d ago 7 comments

Ask HN: Recommandation for an Ergonomic Keyboard?

8 points by cx42net 2d ago 27 comments

Ask HN: Someone has committed 20K+ LoC to a PR, exhausting my CI & AI workflows

4 points by zacian 2d ago 11 comments

GPUPrefixSums – state of the art GPU prefix sum algorithms

49 coffeeaddict1 11 8/28/2025, 12:49:10 PM github.com ↗

Comments (11)

luizfelberti · 29m ago

This looks amazing, I've been shopping for an implementation of this I could play around with for a while now

They mention promising results on Apple Silicon GPUs and even cite the contributions from Vello, but I don't see a Metal implementation in there and the benchmark only shows results from an RTX 2080. Is it safe to assume that they're referring to the WGPU version when talking about M-series chips?

m-schuetz · 1h ago

That and https://github.com/b0nes164/GPUSorting have been a tremendous help for me, since CUB does not nicely work with the Cuda Driver Api. The author is doing amazing work.

coffeeaddict1 · 3h ago

Related paper by the authors: https://dl.acm.org/doi/10.1145/3694906.3743326

genpfault · 4h ago

https://en.wikipedia.org/wiki/Prefix_sum#Applications

almostgotcaught · 4h ago

this is missing the most important one (in today's world): extracting non-zero elements from a sparse vector/matrix

https://developer.nvidia.com/gpugems/gpugems3/part-vi-gpu-co...

merope14 · 3h ago

Not even close. The most important application (in today's world) is radix sort.

WJW · 1h ago

What specific application do you have in mind that radix sort is more important than matrix multiplication?

otherjason · 23m ago

I think they were trying to say “radix sort is a more important application of prefix sum than extraction of values from a sparse matrix/vector is.”

woadwarrior01 · 23m ago

Top K sampling comes to mind, although it's nowhere nearly as important as matmult.

almostgotcaught · 3m ago

ranking models benefit from gpu impls of sort but yup they're not nearly as common/important as spmm/spmv

m-schuetz · 18m ago

Is that relevant for 4x4 multiplications? Because at least for me, radix sort is way more important than multiplying matrices beyond 4x4. E.g. for Gaussian Splatting.