frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Show HN: DualMix128 – A fast, simple PRNG passing PractRand (32TB) and BigCrush

https://github.com/the-othernet/DualMix128
5•the_othernet•1y ago
Hi HN,

I'd like to share DualMix128, a pseudo-random number generator I developed. It's written in C and achieves very high speed while maintaining strong statistical properties for non-cryptographic use.

GitHub (MIT License): https://github.com/the-othernet/DualMix128

Key points:

* *Fast:* Benchmarked at ~0.36 ns per 64-bit generation on GCC 11.4 (-O3 -march=native). This was over 2x faster than `xoroshiro128++` (~0.74 ns) and competitive with `wyrand` (~0.36 ns) in the same tests. (Benchmark code/results in repo).

* *Robust:* Passes the full TestU01 BigCrush suite and PractRand up to 32TB with no anomalies reported. (Test harnesses/logs in repo).

* *Simple Algorithm:* Uses a 128-bit state (two `uint64_t`) with simple mixing based on addition, XOR, rotation, and a final multiplication by the golden ratio constant.

* *C Implementation:* Minimal dependencies (`stdint.h` for core), straightforward C code.

This came out of exploring simple constructions for efficient PRNGs that still hold up under heavy statistical testing.

Curious to hear feedback on the design, performance, potential applications, or any weaknesses you might spot. Thanks!

Comments

thomaskoopman•1y ago
Very cool, fast and looks like it should vectorize too. Do you have a jump function for parallel seeding?

How did you come up with this, some number-theoretic basis or more experimental?

Gloat: Q2 Grant Halfway Report

https://gloathub.org/blog/2026/06/16/gloat-q2-grant-halfway-report/
1•tosh•33s ago•0 comments

GLM 5.2 is now available via a unified Model API

https://www.hpc-ai.com/model-apis
1•hpcaitech•53s ago•1 comments

There Were Lots of Invitations, and I Know You Sent Me Some

https://www.astralcodexten.com/p/waiting-for-the-miracle
2•Michelangelo11•4m ago•0 comments

An URL shortner for all Codeberg users

https://libl.in
1•mindaslab•6m ago•0 comments

South Korea's Fake Online Stores Help Shopping Addicts Save Money

https://www.odditycentral.com/news/south-koreas-fake-online-stores-help-shopping-addicts-save-mon...
2•networked•10m ago•0 comments

CBC will no longer air NHL games in 'end of an era' as broadcast deal expires

https://www.cbc.ca/sports/cbc-nhl-hockey-night-in-canada-ends-9.7236977
3•canucker2016•11m ago•0 comments

Zigzag Decoding with AVX-512

https://zeux.io/2026/06/17/zigzag-decoding-avx512/
1•matt_d•11m ago•0 comments

The 14-point US-Iran peace plan, annotated

https://www.cnn.com/2026/06/17/politics/us-iran-memo-annotated-intl-vis
1•SilverElfin•11m ago•0 comments

Trump admin blocking Fable 5 rerelease unless Anthropic ensures no jailbreaks

https://www.wired.com/story/the-white-house-wants-anthropic-to-block-all-jailbreaks-that-may-not-...
3•reasonableklout•12m ago•1 comments

JPMorgan Chase cuts off Anthropic access for its Hong Kong staff

https://www.ft.com/content/de83d303-6a03-456b-bfb9-7b11dd502ab3
2•frb•12m ago•0 comments

An open-source AI just beat OpenAI's GPT-5.5 at coding (1/6th the price)

https://docs.z.ai/guides/llm/glm-5.2
1•Raj_Sidwadkar•16m ago•0 comments

GLM 5.2 playing text adventures

https://entropicthoughts.com/glm-5-2-playing-text-adventures
1•kqr•17m ago•0 comments

An Open Source Implementation of Notebook LM

https://github.com/lfnovo/open-notebook
1•saikatsg•18m ago•0 comments

Explaining Attention with Program Synthesis

https://arxiv.org/abs/2606.19317
1•E-Reverance•22m ago•0 comments

Data Viz and Table Design from the Letterpress Era

https://chris-parmer.com/data-viz-from-the-letterpress-era/
1•chriddyp•22m ago•0 comments

Estimate and pay your Q2 taxes

https://www.kip-ai.com/estimate
1•prudhvid•23m ago•0 comments

Does Your Programming Language Ever Surprise You in a Good Way?

https://donraab.medium.com/does-your-programming-language-ever-surprise-you-in-a-good-way-eb29e1e...
1•ingve•31m ago•1 comments

The Australian Government to Require SMS/MMS Sender ID Registraion

https://www.acma.gov.au/sms-sender-id-register
12•anitil•33m ago•2 comments

How to Become a Person After Smartphones Have Rotted Your Brain

https://www.thenewatlantis.com/publications/how-to-become-a-person-after-smartphones-have-rotted-...
8•the-mitr•33m ago•0 comments

Six months of AI in 2026, and a whole lot of noise

https://jatniel.dev/en/bytes/six-months-of-ai-in-2026-and-a-whole-lot-of-noise
1•jtnl•35m ago•0 comments

LLMs: Don't use a sledgehammer when tweezers will do

https://superuserdone.com/posts/2026-06-18-dont-use-a-sledgehammer/
1•SuperUserDone•36m ago•0 comments

Catching bad LLM behavior: OpenAI's new Deployment Simulation

https://openai.com/index/deployment-simulation/
1•Raj_Sidwadkar•37m ago•0 comments

DeepSeek Introduces Vision

https://chat.deepseek.com/
14•RIshabh235•39m ago•0 comments

Tesla remotely disabled the €6,200 FSD feature I had used for months

https://teslatookwhatipaidfor.com/en/
3•bartoszhernas•40m ago•4 comments

Environment AI writing code for simulations to test new models of particles

https://github.com/openwave-labs/openwave/blob/main/MODELS.md
1•eln1•44m ago•1 comments

Stack Overflow for Agents

https://stackoverflow.blog/2026/06/10/announcing-stack-overflow-for-agents/
3•marinesebastian•48m ago•0 comments

World-famous Robin Hood oak tree has died

https://news.sky.com/story/world-famous-robin-hood-oak-tree-has-died-13555186
3•austinallegro•1h ago•0 comments

Cervical cancer deaths fall to zero in young women given vaccine

https://www.bbc.com/news/articles/c621z28z138o
4•geox•1h ago•0 comments

What my trade data taught me that 3 years of reading couldn't

https://www.traderscompanion.org
1•mathono•1h ago•1 comments

MBook – a proposal for a new, simple e-book format based on Markdown

https://kevinboone.me/mbook.html
2•mindracer•1h ago•0 comments