DragonFlyBSD Kernel Gets Major SMP Improvements

[phoronix]

Phoronix: DragonFlyBSD Kernel Gets Major SMP Improvements

The latest multi-threading improvements to the DragonFlyBSD kernel involve reducing SMP contention for allowing the BSD operating system to better utilize large multi-core systems...

http://www.phoronix.com/vr.php?view=MTQ4ODU

[Guest]

Good news. This is an only BSD distribution which makes something on its own ie. HAMMER file system. It's also worth to mention its developers don't play in cheap and stupid propaganda lies like those from freebsd.

[dnebdal]

Phoronix: DragonFlyBSD Kernel Gets Major SMP Improvements

The latest multi-threading improvements to the DragonFlyBSD kernel involve reducing SMP contention for allowing the BSD operating system to better utilize large multi-core systems...

http://www.phoronix.com/vr.php?view=MTQ4ODU

Nice - now that their cleanup projects are beginning to solidify, I wonder if they'll finally meet their goal of scaling better than FreeBSD?
(Not that I'll be moving right off ... I'm really quite fond of ZFS, and the things they've done with pkg/poudriere lately are rather useful.)

[ryao]

Nice - now that their cleanup projects are beginning to solidify, I wonder if they'll finally meet their goal of scaling better than FreeBSD?
(Not that I'll be moving right off ... I'm really quite fond of ZFS, and the things they've done with pkg/poudriere lately are rather useful.)

They could always port ZFS.

[Guest]

Nice - now that their cleanup projects are beginning to solidify, I wonder if they'll finally meet their goal of scaling better than FreeBSD?
(Not that I'll be moving right off ... I'm really quite fond of ZFS, and the things they've done with pkg/poudriere lately are rather useful.)

How do you know DragonflyBSD scale worse? As far as I remember it was better than freebsd at Phoronix last time.

[LightBit]

Look at want they are doing, they are just trying to make their OS like Linux. Why don't they just move to Linux instead make it better rather then wasting time and developing a rival knockoff of Linux and fragmenting the open source community?

They are not tryng to make it like Linux.

There's nothing good about ZFS, it's old, out dated, inadequate for modern use and BTRFS has shown to beat it in everything.

BTRFS is not even stable yet.

pkgng is a complete mess which shows just how pathetic FreeBSD is at implementing a simple decent package manager even when it doesn't approach the capabilities of apt-get and yum.

I thought pkg and ports are decent package managers.

[JX8p]

Doesn't matter, Linux still scales way better then both BSDs. Also, DragonflyBSD and community may still be slightly better then FreeBSD but they are still full of crap.

Look at want they are doing, they are just trying to make their OS like Linux. Why don't they just move to Linux instead make it better rather then wasting time and developing a rival knockoff of Linux and fragmenting the open source community?



There's nothing good about ZFS, it's old, out dated, inadequate for modern use and BTRFS has shown to beat it in everything.

pkgng is a complete mess which shows just how pathetic FreeBSD is at implementing a simple decent package manager even when it doesn't approach the capabilities of apt-get and yum.

1. pkgng still does not have an official repository

2. It's slower then all other Linux package manager

3. If you have two different versions of a package (i.e. vim and vim-gtk) installed pkgng doesn't work reporting a package conflict and does nothing.

4. Point no.3 is so annoying because it often happens during upgrading of packages. There be package-2.0 install needing to be replaced by package-3.0. pkgng sees this and stops giving an error saying package-3.0 conflicts package-2.0 and now you are stuck.

poudriere is nothing more then a really pathetic way to make up for the lack of a pkgng repository. Now users have to make thier own repository (aka. compile the damn entire ports tree again and again wasting so much time).

DragonFly BSD are not trying to copy Linux. Matt Dillon, the main developer, is more inspired by AMiGA. Furthermore, there are many reasons why people prefer to develop for Dragonfly than for Linux. I will leave the research up to you.

ZFS is certainly getting older, but it is still leaps and bounds ahead of most filesystems used today.

pkgng is not a mess as far as I can tell. It works just as well as any other package manager I've seen, and I would say that it definitely closely matches the capabilities of apt and yum.

pkgng does not yet have an official repository. This is correct. You shall have to wait for one. Alternative, there a wide variety of repositories by third parties which would work just as well.

I've not seen it to be any slower than most other package managers, personally. Are you sure?

Your 'points' number 3 and 4 are things I've never heard of-- perhaps if you experience those you might wish to file a bug report and try to resolve the issue?

As for poudriere, I quite like it, personally. It is not a replacement for a central package repository but it is very useful regardless. Particularly in that many of us prefer to customise and configure our packages before building.

[enjolras]

pkgng has no official repo for FreeBSD, but there has been official repos for DragonFly for quite some time now, with more than 20k pre-built binary packages avaible.

[ryao]

ZFS is certainly getting older, but it is still leaps and bounds ahead of most filesystems used today.

ZFS is under active development and continues to improve upon itself. While other filesystems are still behind what ZFS was in Open Solaris, ZFS itself is well ahead of that. There is a list of notable changes/improvements maintained on the Open ZFS website:

http://open-zfs.org/wiki/Features

[intellivision]

Why don't they just move to Linux instead make it better rather then wasting time and developing a rival knockoff Linux and fragmenting the open source community?

Whoops, best ask Google to stop making Android and Samsung to stop developing Tizen then. Can't have any fragmentation at all /s