[ I’m breaking from my usual technical posts, since I have some wonderful news to share with you all; opinions mine, not those of my employer, etc. ] For readers of elevated left-leaning publications such as the Guardian, one of the more remarkable discoveries one might make is that the sources of environmental devastation are… Continue reading The Environment will Be Saved if We Just Avoid Doing Icky Things→| Branch Free
A quick note – while our main literal matcher and overall approach are described in our earlier paper (link and discussion https://branchfree.org/2019/02/28/paper-hyperscan-a-fast-multi-pattern-regex-matcher-for-modern-cpus/), we didn’t have space to discuss in much detail other string matching approaches. Our “FDR” matcher is used for large-scale literal string matching – but what if the number of literals is small?… Continue reading Hyperscan Papers: Harry and Teddy (SIMD litera...| Branch Free
I’ve made many different types of SIMD usage over the years, and occasionally even written about them (on this blog, in papers, and in Twitter). Just for reference, I’d like to cover my taxonomy of SIMD usage so I don’t keep writing it up (poorly) on Twitter every few months. This discussion is a draft… Continue reading A (Draft) Taxonomy of SIMD Usage→| Branch Free
With Computex, there’s been a ton of news about Ice Lake (hereafter ICL) and the Sunny Cove core (SNC). Wikichip, Extremetech and Anandtech among many others have coverage (and there will be a lot more), so I won’t rehash this. I also don’t want to embark on a long discussion about 14nm vs 10nm, Intel’s… Continue reading Why Ice Lake is Important (a bit-basher’s perspective)→| Branch Free
I’m pleased to report that Hyperscan, the regular expression matcher that ate my life from 2008 through 2018, finally has a paper (pdf) – it’s being presented this week at NSDI &#…| Branch Free
There is a persistent meme out there that matching regular expressions with back-references is NP-Hard. There is a post about this and the claim is repeated by Russ Cox so this is now part of received wisdom. None of these claims are false; they just don’t apply to regular expression matching in the sense that… Continue reading Question: Is matching fixed regexes with Back-references in P?→| Branch Free
(the author hard at work) In the last post, I talked about some first impressions of programming SIMD for the ARM architecture. Since then, I’ve gotten simdjson working on our ARM box – a 3.3Ghz eMag from Ampere Computing. I will post some very preliminary performance results for that shortly, but I imagine that will… Continue reading Fitting My Head Through The ARM Holes or: Two Sequences to Substitute for the Missing PMOVMSKB Instruction on ARM NEON→| Branch Free
(the pictured dish is apparently materials for “Buddha Jumps Over The Wall”, named for its ability to seduce away vegetarians – sadly it uses shark fin so has some ethical issues…) [ UPDATE: I have, at least partly, dealt with the lack of PMOVMKSB and written a new post about it ] I’ve done a lot… Continue reading An Intel Programmer Jumps Over the Wall: First Impressions of ARM SIMD Programming→| Branch Free
Well, if that doesn’t put off some readers, I don’t know what will… A technique that we recently disclosed for our JSON parser (see the paper or recent blog post) is something I think I invented (please set me straight if this is known since the glory days of HAKMEM, or whatever). We want to… Continue reading Code Fragment: Finding quote pairs with carry-less multiply (PCLMULQDQ)→| Branch Free
Daniel Lemire and I have spent some time this year working on a fast JSON parser. You can read about it in paper form at arXiv here and the repo is here. In this blog post, I’ll provide an informal summary of the paper and some background as to the thinking behind the system. What… Continue reading Paper: Parsing Gigabytes of JSON per Second→| Branch Free