MultiQueues: Simpler, Faster, and Better Relaxed Concurrent Priority Queues | Hacker News Reader

MultiQueues: Simpler, Faster, and Better Relaxed Concurrent Priority Queues

arxiv.org

46 pointsuser_23571110 comments

The SprayList paper is quite a bit more detailed, and it's underhanded how the MultiQueues paper shoves the bad news into an addendum, but the interesting thing to me is these papers coming out within weeks of each other.

The SprayList: A Scalable Relaxed Priority Queue http://research.microsoft.com/apps/pubs/default.aspx?id=2091...

Unfortunately Microsoft patents at least some of the MSR research on algorithms and data structures, which makes it effectively unusable for many users.

Can you provide more information? How do you determine which paper is patented?

I only skimmed it, but get the idea it's similar to this: http://www.boyet.com/Articles/LockFreeLimitedPriorityQ.html

This kinda reminds me of the opening scenes of Silicon Valley.

Just imagine someone getting up on stage and yelling enthusiastically, "Simpler, faster, and better relaxed concurrent priority queues!" To a round of muted applause.

I'm sure this submission has tons of technical merit and all that. But without context, that scene from Silicon Valley is all I can really think of. :)

The context is the paper. Please try not to read a title and post an inane response.

Context should be established by headlines on news sites like Hackers News.

I know we're a technical audience, but I'd be willing to wager more than half of this site's audience has no clue what that title is talking about.

Prefixing it with something like "A faster algorithm for [purpose]" is helpful. Otherwise it can just sound like technobabble.

I don't think it's inane to point out that jargon sounds jargony to those not in the know.

Probably, the whole paper is written in that terse tone. If the title doesn't appeal to the reader, reading on wouldn't make sense anyway. I.e., I can tell from the title that I'd be better of returning to the other book on data structures.

I'll download it for later(tm).

Priority queues are an incredibly important concurrency primitive. This is very important work.

If you weren't expecting to find detailed technical articles, you might be on the wrong website.

Reply on news.ycombinator.com