DragonFlyBSD 3.8 Will Be Released Soon

[phoronix]

Phoronix: DragonFlyBSD 3.8 Will Be Released Soon

The plan is to release DragonFlyBSD 3.8 in early June...

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

[drSeehas]

"Benchmarks will come in short order after the official DragonFlyBSD 3.8 release."
It will be nice, if you compare it to FreeBSD.

[BSDude]

I'm really curious about the Hammer2fs. How it will stack against ZFS or BTRFS for that matter?! Although it will be waaay behind both filesystems in regards of stability and feature parity it can be a great alternative. My only gripe with ZFS is the fact that a wrapper is needed for system calls.

[name99]

"Benchmarks will come in short order after the official DragonFlyBSD 3.8 release."
It will be nice, if you compare it to FreeBSD.

Such a comparison is not really that helpful. Remember that DragonFly is not targeted at the same sort of machines and markets as FreeBSD --- it is designed for really large systems, and all choices have been made to allow for scaling to very large machines.
That means that if you benchmark it on small machines, chances are you're going to get it appearing to be slower than FreeBSD. This is no surprise, and not upsetting to any member of the DragonFly team.

Basically, it's like comparing a kernel designed and built for a mono-processor to a kernel built for multi-processing. Yeah, the multi-processing kernel is slower if the machines you care about are single processor. But the single processor kernel is not especially useful if what you care about is an MP machine...

[drSeehas]

Such a comparison is not really that helpful. Remember that DragonFly is not targeted at the same sort of machines and markets as FreeBSD --- it is designed for really large systems, and all choices have been made to allow for scaling to very large machines.
That means that if you benchmark it on small machines, chances are you're going to get it appearing to be slower than ...

... everything else at phoronix.
So you want to say: Please don't benchmark Dragonfly at all unless you have a really large system?

[endman]

Remember that DragonFly has been made to allow for scaling to very large machines.

Citation needed....

All BSDs need very large and powerful machines to achieve 50% the speed of Linux running on a netbook. That's a known fact.

Basically, it's like comparing a kernel designed and built for a mono-processor to a kernel built for multi-processing. Yeah, the multi-processing kernel is slower if the machines you care about are single processor. But the single processor kernel is not especially useful if what you care about is an MP machine...

No, it's a fair comparison. But to make a more sensible benchmark, include Linux in the comparison.

[name99]

Citation needed.... [that target of DragonFly is scalability]

Oh for crying out loud. That's like asking for proof that the sky is blue. Reading anything by Matt Dillon makes this obvious.

Go to the damn Dragonfly page. Let's look at eg:

http://www.dragonflybsd.org/history/

"The ultimate goal of the DragonFly project at its inception was to provide native clustering support in the kernel. This type of functionality requires a sophisticated cache management framework for filesystem namespaces, file spaces and VM spaces. ... This original design direction, although no longer the primary goal of the DragonFly BSD project, has influenced many of the design decisions made in the intervening years."

OR

"Recently, many developers have focused on SMP scalability"

OR

http://www.dragonflybsd.org/performance/

"the project sought to do SMP in more straightforward, composable, understandable and algorithmically superior ways to the work being done in other operating system kernels"

[Vim_User]

Don't bother with endman, he is just another sock puppet of the anti-BSD troll. Just put him on your ignore list, nothing intelligent or true will come from that one anyways.

[endman]

Oh for crying out loud. That's like asking for proof that the sky is blue. Reading anything by Matt Dillon makes this obvious.

Go to the damn Dragonfly page. Let's look at eg:

http://www.dragonflybsd.org/history/

No one who's serious about building a working system would trust Mad Dog Dillon anymore. Too many lies. Just look at Hammerfs 2, Mad Dog Dillon promise a release for 2012. Yet today in 2014, it's still vaporware (source code can even be found).

"The ultimate goal of the DragonFly project at its inception was to provide native clustering support in the kernel. This type of functionality requires a sophisticated cache management framework for filesystem namespaces, file spaces and VM spaces. ... This original design direction, although no longer the primary goal of the DragonFly BSD project, has influenced many of the design decisions made in the intervening years."

OR

"Recently, many developers have focused on SMP scalability"

OR

http://www.dragonflybsd.org/performance/

"the project sought to do SMP in more straightforward, composable, understandable and algorithmically superior ways to the work being done in other operating system kernels"

That was the original goal of DragonflyBSD. They have since abandoned it because of their incapability to compete with Linux (just like other BSDs with their goals). The project now has no clear projects or road maps just like other BSDs. So BSD fanboy should stop dreaming and fact the real world.

Linux is the future.

Don't bother with endman, he is just another sock puppet of the anti-BSD troll. Just put him on your ignore list, nothing intelligent or true will come from that one anyways.

Citation needed.....