FreeBSD Release Engg. Team’s Murray Stokely has put together a 4.9-BETA ISO image for i386 with the latest package build.
FreeBSD 4.9-BETA i386 ISO with Packages Available Now
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
… about programming asks:
Why is world + dog still reffering to i386 (at least name-wise)..?!! W2K install-directory does as well. Certainly noone would attempt to install either W2K or BSD on 386 — so where is the link between current hardware and this i386-thinggy I am missing…
Today’s Pentiums aren’t a lot different to 386s from a programmer point of view. They may have changed drastically inside, but the ISA is basically the same with extra add-ons (MMX, SSE). The way to program memory protection, paging, task switching, basic floating point, etc. hasn’t changed since the 386.
Maybe IA-32 would be a better name?
i386 is just another way of referring to IA32. The 386 was the first 32-bit processor in the x86 family, and therefore the absolute minimum processor required to run a 32-bit operating system on x86.
because when it comes to Gentoo (?), others, people seem to freak about why it isn’t compiled for 686 instead 586 and alike… allegedly missing the features, performance, etc…
I don’t think an OS would benefit much from compiling for i686, unless you’ve got a driver that does a lot of FP calculations. I’m of the opinion that such code shouldn’t be in a kernel, anyway 😉
because when it comes to Gentoo (?), others, people seem to freak about why it isn’t compiled for 686 instead 586 and alike… allegedly missing the features, performance, etc.
FreeBSD is compiled with i686 optimizations by default
Isn’t x86 the best name for the Intel arch. ?
I want to try BSD (FreeBSD) on my laptop (i830M based). Any experience on this hardware? Is ACPI supported in BSD? The handbook says that the 5.x supports it, but I don’t want to reinstall to find out…
Ironically, GCC3.2 isn’t able to create TRUE i386 code, so I dont even think you could run modern versions of FreeBSD on a 386 if you wanted to.
Slackware had this problem, thats why Pat changed to i486.
“Is ACPI supported in BSD?”
It is enabled by default in FreeBSD 5.x, and it will be available as an option on FreeBSD 4.9. It’s been my experience that the FreeBSD ACPI code is considerably more stable than Linux’s implementation.
Another really cool thing that is now default in FreeBSD 5.x is devfs. Man I love that…
This is just my luck. Sunday, I decided to revamp one of my older machines. I then decided to make it dual boot FreeBSD 4.8 and BeOS Max v3.0 beta3. Ever get the feeling that you are being toyed with?
It is enabled by default in FreeBSD 5.x, and it will be available as an option on FreeBSD 4.9. It’s been my experience that the FreeBSD ACPI code is considerably more stable than Linux’s implementation.
Linux’s ACPI implementation now is where FreeBSD’s was two years ago.
FreeBSD is able to throttle down idle CPUs automatically, greatly reducing energy consumption and heat when the system is idle. This was a godsend for my SMP system at home, which is a small space heater otherwise.
Does anybody know if SIS 900 PCI ethernet adpter support is included in the 4.9?
http://www.sideliners.ca
Freelance Empowermnent
//
FreeBSD is able to throttle down idle CPUs automatically, greatly reducing energy consumption and heat when the system is idle. This was a godsend for my SMP system at home, which is a small space heater otherwise.
//
Actually, this is a new feature in the 2.6 kernel. It works well (though I dont know if it works better than BSD).
Actually, this is a new feature in the 2.6 kernel. It works well (though I dont know if it works better than BSD).
I have Linux 2.6.0-test5 on the same system. Either I missed the kernel option for this or it doesn’t work automatically.
Great, this will improve my battery-time… about 2 hours total i think (small battery, small laptop), but what about suspend and standby then? In linux i use the 2.4.x series, and .20 and .21 with acpi patches… I managed to get the poweroff, and lid switch working, and default mandrake install even provided a “acpi” tool which gave me the system1 and system2 temperature on the system. But I have never got it (under linux) to suspend and/or standby… This would give me even more work-time
(sorry for bad English)
Still trying to get my audigy working under 5.1 and xft2. I am a novice. Everything else works great. Just no sound and the fonts looks terrible.
Any help would be appreciated or point me the right direction to look for help.
Thanks.
I really can’t help with you on the sound, but did you add option (device pcm) in the kernel or load it as module (/boot/loader.conf) yet? You can check in /boot/defaults/loader.conf for the hint.
As for font, you should make sure you have those in the /etc/X11/XF86Config following:
Load “freetype”
Load “type1”
Also, install x11-fonts/bitstream-vera, it’s very nice fonts. Make sure to run ‘fc-cache -f -v’ as root when you add some fonts.
what does ‘fc-cache -f -v’ do? Why do i have to type that when adding fonts?
http://chibis.persons.gfk.ru/audigy/
Check out that site. You’ll need to download his latest driver and a patch, but it works great for my Audigy.
Adam
what does ‘fc-cache -f -v’ do? Why do i have to type that when adding fonts?
Well, sometimes you will need to run ‘fc-cache -f -v’ to teach fontconfig about those new fonts. Sometimes, you don’t need to do that and done by automatic. Fontconfig/Xft2 can be weird sometimes.
May I be wrong, there are years by now that I don’t program in assembly but, as far as I remember, we have some penalties in memory unaligned access and also some optimizations in copying operations and other stuff too.
In this way, can really make sense compile for a little “modern” processor like i686.
I use to ever customize the kernel of machines where I install Slackware or FreeBSD and set the default processor to be i686 (or better when possible) for any new compilation.
I didn’t conducted a benchmark but that procedure give me the feeling of a little faster system.
I have 5.1_RELEASE_p5 on my Athlon XP 2000+ system (Asus A7n266-vm nforce mobo). The boot messages told me ACPI wasn’t working (and, in fact, wasn’t), so I disabled it in the boot options.
I was wondering if anyone knows why it doesn’t work. As far as I can tell, the bios allows for it and all power options work in Windows.
Kingston wrote:
It is enabled by default in FreeBSD 5.x, and it will be available as an option on FreeBSD 4.9. It’s been my experience that the FreeBSD ACPI code is considerably more stable than Linux’s implementation.
Bascule wrote:
Linux’s ACPI implementation now is where FreeBSD’s was two years ago.
Ehe, allow me to quote from Scott Long’s 5-STABLE Roadmap (http://kerneltrap.org/node/view/585):
– ACPI…Linux uses the same Intel vendor sources as FreeBSD, so we should investigate how they have handled some of the known problems.
And if you check out the acpi-devel mailing list or the changelog for that code you will notice that several *BSD-folks (NetBSD uses the same code) are active contributors. The code is dual-licensed.
Well aware of that. The difference seems to be in how actively each side is working on it.
Hi, I am now using slackware as my production server. But I am considering FreeBSD as an alternative.
My division uses Oracle as a backend DB server. I know there is no native Oracle for FreeBSD, to run it under FreeBSD we have to use Linux simulation.
I want to know how well oracle ran under linux simulation. Is it reliable for production server? Is there any posible problem using Linux simulation for any kind of application?
And considering that my server has 2 CPUs, is its performance really scalable (exp. 1 cpu=10, 2 cpu= 20) like linux kernel.
Thanks for your enlightenment.
With some programs, it’s possible to run them w/ FreeBSD Linux emulation even faster than on Linux itself.
You should read FreeBSD handbook -> Linux emulation.
Well aware of that.
I honestly doubt that.
The difference seems to be in how actively each side is working on it.
If you’re comparing a 2.4-kernel from kernel.org to FreeBSD 5.x, sure. I believe the commercial vendors have been a lot more aggressive about adding ACPI-support to their kernels though. A lot of ACPI stuff went into 2.4.22 as well.
With some programs, it’s possible to run them w/ FreeBSD Linux emulation even faster than on Linux itself.
Speed is not my main concern. If we want fast server we will add memory or replace with faster processor.
The main question is still, is it reliable?
Can we feel secure with it? I need real life experiences.
Thanks.
For me, Acrobat Reader as provided by FreeBSD works fine, though the French StarOffice 5.2 installer just coredumps before starting up.
It might be that I have not installed Linux versions of certain libraries, but still.
But can’t you just try it out first on a test server or so?
“I honestly doubt that.”
Check out some of my prior posts on older stories. I damn well am aware of it.
“If you’re comparing a 2.4-kernel from kernel.org to FreeBSD 5.x, sure.”
The 2.6 code is much improved, but it is still very flakey.
“A lot of ACPI stuff went into 2.4.22 as well.”
And did they ever do a shoddy job of it. True FreeBSD’s current version doesn’t work on all systems either, but in general they seem to be making better progress.
Check out some of my prior posts on older stories. I damn well am aware of it.
Why didn’t you just provide me with a link to such a comment?
And did they ever do a shoddy job of it.
Indeed? a) the code is pushed to Linux by the very people who maintain it, iow the very same people who have largely written FreeBSD’s ACPI implementation. b) the patches prepared by these people has had long exposure in the kernels shipped from several of the commercial vendors.
That said, from what I’ve read there are some very real problems with the ACPI code. The Intel engineers who write most of the code are very set on sticking strictly to Intel’s spec. The problem is that a lot of hardware that claims to be ACPI compliant isn’t, and from what I’ve read on LKML the Intel folks haven’t been to good at including fixes for screwy hw. As one who regularly skims through several FreeBSD mailing lists (including -current), I _know_ there are a lot of problems with it. Heck, the roadmap even mentions ACPI as a big issue that needs to be straightened out. But hey, I guess the Linux part of it is just magically shoddy, while the FreeBSD code is two years ahead.
Also, allow me to make another little point. I _ALWAYS_ try to back up anything I write that isn’t purely an opinion. And even when I read something that I _know_ is wrong (which is much too often on these pages), if I either can’t find what I’m looking for to back it up because I can’t remember exactly where or when I saw it or I’m just not in the mood to dig through mailing lists that night…do you know what I do? I DON’T POST! Because random claims with nothing to back them up are worth zilch. Atleast to me.
Obviously, that’s not how you look at things. It’s shoddy, Kingston has spoken.
The 2.6 code is much improved, but it is still very flakey.
True FreeBSD’s current version doesn’t work on all systems either, but in general they seem to be making better progress.
It must be magic. In FreeBSD the code ‘just works’, while in the Linux kernel the very same code turns to mush. Makes sense.
It might be that I have not installed Linux versions of certain libraries, but still.
But can’t you just try it out first on a test server or so?
Yes, I will.
I just want to listen from others to compare the result with my test. I want to know the whole picture not just my experience, if any difference result come up.
You are right, I sometimes post when I should not, or neglect to provide links to back up my claims. I will make more of an effort to back up my claims in the future, and make it clear when I am giving my opinion.
Thanks for the (needed) kick in the ass. My appologies.
In response to the last part of your last post though, In my experience, on a number of different machines, the FreeBSD implementation of ACPI does in fact “just work” whereas that Linux one does not.
I have run into situations where Linux “just works” where BSD does not as well, but ACPI support was never one of those.