Hacker News Reader

Comment by Grollicus | Hacker News Reader

The Speedup numbers named in the commit [0] are impressive on arm but don't look very well on x86. On the other hand they add a thread-local write to every context switch and add a bunch of code. Add to that only 1.2 speedup on the LTTng benchmark.

To me as a non-kernel guy this doesn't sound very impressive..

[0] https://github.com/torvalds/linux/commit/d7822b1e24f2df5df98...

View on HN

Full disclosure, I work on LTTng and a 20% improvement is a huge deal to our users, so I'm certainly biased here.

As far as I know, the thread-local write only occurs if a rseq region is set. The added cost to a context switch seems essentially limited to a branch: https://github.com/torvalds/linux/commit/d7822b1e24f2df5df98...

The other benchmarks that were posted in the various threads leading to this merge looked fairly impressive to me.

Of the top of my head, jemalloc benefited a lot from this patchset, and not just on the run-time front: https://lkml.org/lkml/2016/10/10/332

Ah yes, the ubiquitous middlebrow dismissal (https://news.ycombinator.com/item?id=4693920 http://www.byrnehobart.com/blog/why-are-middlebrow-dismissal...). "As a non-kernal guy" gives it away immediately.

The other responses were useful. Yours is just a put-down. Why bother? To discourage "middlebrow" comments? But the one you're responding to elicited useful information, so it's hardly a problem.

I found the articles on the "middlebrow dismissal" insightful. And I don't think ends usually justify means; I think the parent-most comment was not as nice as it should have been.

The parent comment is not so much a put-down as much as it calls out a particularly unsophisticated dismissal. The parent-most comment would have elicited the same useful information if it were phrased as a question, but would have been less uncouth. I think that's a perfectly fine thing to write a comment about.

A 1.2x speedup affecting a primitive operation is huge.

Reply on news.ycombinator.com