I’m faster than you! No I am!

All FreeBSD interested people will remember the document that Kris Kennaway released (Introducing FreeBSD). In this paper he explains how dramatic improvements have been achieved in FreeBSD 7.0; especially with regards to SMP and SQL database querying (MySQL and PostgreSQL). According to his findings FreeBSD even outperformes Linux.

There’s always been a healthy competition between Linux and FreeBSD, but stating that FreeBSD is faster than Linux, that hurts….

After major improvements in SMP support in FreeBSD 7.0, benchmarks show it performing 15% better than the latest Linux kernels (PDF, see slides 17 to 19) on 8 CPUs under PostgreSQL and MySQL. While a couple of benchmarks are not conclusive evidence, it can be assumed that FreeBSD will once again be a serious performance contender.

Linux kernel developer Nick Piggin reran the benchmark and came to a different conclusion: In his benchmark Linux was faster than FreeBSD.

I’m not an expert, but what do you guys think of this? Is Nick doing a fair analysis and comparison? Anybody been using FreeBSD 7 in a “heavy duty” (SQL) environment who can comment on this?

8 thoughts on “I’m faster than you! No I am!

  1. Oliver says:

    >In other words, I can’t say definitively that Linux is faster than FreeBSD. My primary interest is to see that Linux’s performance problems on this

    workload are under control.

    So it’s not faster than you, but it’s okay.

    And according to Jeff Roberson

    (SCHED_ULE),

    http://jeffr-tech.livejournal.com/18706.html

    >I’m also happy to see some collaboration and competition between linux and bsd kernel

    developers. I hope that continues. We’re really more alike than we are different.

  2. Eric says:

    To have

    initial testing proving otherwise, and having a conflicting outcome after the fact, I think that Linux users feels threatened, I’d have to say that Nick was

    not even doing the tests on the same machine or used faster hardware in his box for the Linux tests. The PDF Preview specifically tested 2.6.22.
    The only

    other reason is that Nick gave the server treads real-time scheduler priority on Linux and not actually let the kernel manage the requests. That’s my take,

    I haven’t had any chances for any heavy duty SQL environment testing, but I just think the Linux side of the fence feels threatened lol. I use them both,

    and Linux feels faster only on desktop responsiveness not actual task execution, but for server and development, nothing has beaten FreeBSD for this task in

    my home personal use regiment.

    That’s all for now :P, Later.

  3. kace says:

    If you look at the graph that Nick has (helpfully) linked to (http://people.freebsd.org/~jeff/sysbench.png), then it seems that his results are

    reasonable and consistent with prior data given that 1) linux has corrected its scaling issues at >8 threads, and 2) Nick tested against SCHED_ULE (not

    SCHED_SMP, which was performing better across the board in Jeff’s graph).

  4. Oliver says:

    >I just think the Linux side of the fence feels threatened lol.

    Do you think so? It seems most of them don’t even now FreeBSD is somehow

    comparable.

  5. Eric says:

    One quick

    fact, people fear what they do not know, so yes, what you say is exactly what I indirectly mean ;) Anything that proves to be better without their knowledge

    is a threat to most Windows converts from what I’ve been seeing.

    Hopefully they will at least try it sometime through PC-BSD or DesktopBSD :)

  6. kace says:

    I think the benchmarking game is a dangerous one to play. I find FreeBSD superior to Linux in so many ways, that I could frankly care less if

    it’s 3.742% behind in this or that benchmark. What about stability and documentation and on and on? (Human time is more valuable that CPU time!)

    That’s

    not to ignore the importance of performance, which is also excellent; or the great progress made by the Project in the SMP schedulers. I’m just saying that

    this story and benchmarks in general sometimes seem to make Linux the standard and performance benchmarks the only yardstick. … We should all know

    better.

  7. Pingback: In a nutshell: 2008-03-08 | adminlife.net

  8. PMurphy says:

    I have to agree with Kace, that a benchmark should not be a measure of the OS’ quality, usability and stabilty. You’ll find that some people spend hours tweaking for better numbers due to some competitive quirk in their nature but add no value to the OS other then the draw factor. Hence the popularity of Linux.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>