FreeBSD 4.10-RC2 is available for download. There are only a few changes since RC1, but it does include the release package set, so feel free to test things like KDE and Gnome. Due to some late issues that are cropping up, it will probably be an RC3 before the final release.
FreeBSD 4.10-RC2 Released
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
This has probably been asked before.
But which version should a new to BSD user install?
The 4.x line or 5.x ?
This has probably been asked before.
It has.
But which version should a new to BSD user install?
The 4.x line or 5.x ?
Production servers should probably go with 4.x and desktops should go with whatever you like more, new features, or stability.
Thanks for the reply.
I have a slow k6-2 450Mhz, would a 4.x be faster than 5.x for desktop use?
I have a slow k6-2 450Mhz, would a 4.x be faster than 5.x for desktop use?
how to qualify desktop usage? production or stability?
has someone already told you go with whatever you feel better. Despite inst the OS that makes the desktop but what you are willing to install on it.
Rule of thumb, software doesnt make miracles, and cant masquerade hardware weakness, but despite that you have a good hardware combo for X usage, not for a ramkiller desktop enviroment like kde or gnome, but a good candidate for xfce or open/black/fluxbox windows environment.
I have a slow k6-2 450Mhz, would a 4.x be faster than 5.x for desktop use?
NO
if you can get away with it use xfce. *box is just too basic. I set it up on an old PowerBook (G3 333 mHz 64Mb) since I wanted the simplest WM I could find but it’s really lacking. I’ve used XFCE on the Freesbie CD and its decent. The biggest problem was finding a filemanager and Gentoo http://www.obsession.se/gentoo/ is the best I could find. I’m still looking for a decent text editor that’ll handle UNICODE (I need french- I learned the hard way that jEdit is not an option)
Now, personally I use nothing but KDE; but then again I’ve got a 1GB RAM and wouldn’t trade KDE for the world.
Which of the two has been shown to work better on older hardware?
well ask what you mean.
4.x is better for older hardware usually
FreeBSD 5.x not only has better support for newer hardware but it also has new features. I’d say go with 5.x and if you find it acting stably, there’s no need for you to go back to 4.x. The 5.x branch shows where FreeBSD is heading.
for the sake of sanity and following someone questioning, gentoo is nice, you still have the xfce file manager, rox or the classic mc (or the best norton commander ever for nixes, xnc)
And I dont find wm like openbox flux or waimeas/kahakais basic, since they follow freebsd’s spirit, you just install whatever you want, whatever you need, nothing more, nothing less and not a bundle of apps that some you will never use on your entire life.
For the thread starter, if you want a number 4 or 5, go 4.9-Release and move up to 4.10-RC2 by cvsuping your src tree to the latest -STABLE branch.
(check the handbook on the way to achieve that)
links:
rox.sf.net
http://www.ibiblio.org/mc/
xnc.jinr.ru
icculus.org/openbox/
kahakai.sf.net
<italic>
“I’m still looking for a decent text editor that’ll handle UNICODE”
</italic>
use vim
http://www.vim.org
I wasn’t too impressed with Linux SMP support so I did some research on how FreeBSD would fair. According to some benchmarks, the 4.x series performs slightly better than 5.x in SMP configurations.
To answer the question about which would be better to run on a slow (ie 450 MHz computer) it’s almost negligable (sp?). What *will* make the difference is the software you run on it, meaning: pick a lightweight desktop, and chosing Firefox over plain Mozilla. Decisions like that will make a bigger impact on your desktop experience, more so than choosing 4.x or 5.x
– According to some benchmarks, the 4.x series
– performs slightly better than 5.x in SMP configurations.
That is surprising since one of the main draws of the 5.x branch is SMPng (FreeBSD’s new fine-grained locking/threading technology).
Were there any apparent reasons for the 5.x performance penalty?
Could you share a link to the stats?
And how does Dragonfly compare to this?
DragonflyBSD is still in the initial developement stages.
Well since this thred is almost 99% OT I’ll clarify that that the editor needs to be graphical for a barely computer literate user.
On a similar note I would have used FreeBSD on that laptop but suprising FreeBSD does not support PPC. PPC is probably the second most common processor out there and definitely the second most common processor on the desktop.
Then again the traditional stronghold of FreeBSD has been servers (and workstations?) where Alpha and Sparc have been around forever. But then again the is no POWER port either…
Anyways PPC suport is on it’s way FreeBSD/PowerPC currently is on the verge of booting to single-user mode.… I wonder how long it will take for them to port it to PPC64 once it’s done.
Anyways…
</rant>
Besides dragonfly’s own site
http://www.dragonflybsd.org/
This article gives a summary
http://www.bsdnews.org/03/dragonfly.php
Great FreeBSD Team….
The PPC port is advancing at a steady pace.
FreeBSD actually does boot to multi-user on PPC, and has done so for more than a year. The biggest problems right now are that it’s kinda unstable under heavy load, and that X doesn’t work. Hopefully, these issues will be fixed before 5.3-RELEASE comes out.
As for ppc64 support, it should come relatively soon too.
On the other hand, the developers don’t have access to POWER hardware, so it might take a while before FreeBSD runs there.
“I wasn’t too impressed with Linux SMP support so I did some research on how FreeBSD would fair”
well. the linux smp support in 2.6 outperforms freebsd 5.x and 4.x. so i dont know what issues you have with it. they both take different approaches to it thou
I ran FreeBSD 5.x on my k6-2/350 for over a year without complaint. Note that both Mozilla and Mozilla Firefox are too dog slow to be usable, especially with more than one tab open — Opera, on the other hand, flies.
I used IceWM as my window manager, but apart from Opera and xpdf (and sometimes mplayer), virtually all software I used was console based (I’d have a maximized borderless rxvt running GNU screen on my main virtual desktops, Opera on my second virtual desktop, and a windowed rxvt on my third). mpg123 with a few scripts for music, vi and nano for text editing, naim for IMing, ssh->pine for e-mail (though sometimes I used Opera instead), etc.
I really liked the performance FreeBSD 5.x offered — it ran far better on my machine than Mandrake/Debian/Slack.
This has probably been asked before.
But which version should a new to BSD user install?
The 4.x line or 5.x ?
Well, I would go for 4.x; I’m on the more conservative side and you’re less likely to see things that are broken in the source tree due to the GCC 3.3 vs. 2.95 differences that exist.
You’ve also got the assurance that only the best tested parts are bought back into the 4.x line.
With that being said, there is nothing wrong per-say about 5.x, however, IMHO I would wait until it has stablised before coming out and celebrating/using it either fulltime as a workstation or desktop.
> well. the linux smp support in 2.6 outperforms freebsd 5.x
> and 4.x. so i dont know what issues you have with it. they
> both take different approaches to it thou
Would you care to qualify that statement? I often experience temporary lock-ups (0.5 – 3 sec) under heavy I/O with Linux 2.6. Neither FreeBSD 5.x and 4.x have ever given me such trouble. With KSE (now present in cvs, due in 5.3), SMP scalability is at least on par with Linux 2.6.
From Matthew Baulch
>I often experience temporary lock-ups (0.5 – 3 sec) under >heavy I/O with Linux 2.6. Neither FreeBSD 5.x and 4.x have >ever given me such trouble. With KSE (now present in cvs, >due in 5.3), SMP scalability is at least on par with Linux >2.6.
I would have to disagree. The potential for SMPng to match SMP in Linux 2.6 is certainly there, but the numbers are still in Linux’s favor and it’s the widely held opinion in freebsd dev that they have quite a bit more ways to go before we (I say that as a long time member of the freebsd user community) have a comparable offering. There’s a lot more giant-freeing work that needs to be done in network subsystems, further stabilizing ULE, revising kqueue, as well as other odds and ends. Not to mention stability and scalability testing of revised code. Your experience with SMP performance may be an exception, but it’s not the rule at this point.
I assume you are somehow affiliated with the project, in that case I say congrats and keep on with the good work. However as there is quitea big difference from on the verge of booting to single-user mode (which is got from the <a href=”http://www.freebsd.org/platforms/ppc.html“>site– though I shoulds have noticed the date on that)and kinda unstable under heavy load in multi-user mode so maybe you could get someone on your team to update the website?
And BTW, not that I own or have been within a 1000m of a POWER box ever, but I assume it couldn’t be that difficult since IIRC PPC = POWER + Motorola 88k. But then again looking now at the NetBSD site they support some PPC but not others so I guess there must be more involved in porting than just the processor…. I guess the rest must be the chipset and expansion card bus?
Whats the newest word on 5-stable from anyone in the know?
Whats the newest word on 5-stable from anyone in the know?
It has been pushed back further. Its a good sign that the FreeBSD developers have realised that it would be best to release a stable version which has had the giant lock completely removed rather than stuffing around release after release. If the giant is removed, then with each minor release (5.x), it will be just a matter of fine graining it further and optimising the code gradually.
IIRC, around July-Sept, however, IMHO, they shouldn’t aim for a date, instead they should concerntrate on getting the things done, and if it means that it has to be pushed back to the end of the year, so be it. Better to release a stable and full featured version than something half-baked.
well work on this should speed up nicely since wasn’t the kernel lock what that guy fund raise to beable to spend the next 6 months doing?