FreeBSD s/390 Port in the works 82
brad-x writes: "It appears that an enterprising gentleman has taken the time to port FreeBSD to the s/390. It needs some work yet, as his project page suggests, but if he makes it happen it will definitely be very cool. Check it out!"
Does this mean that... (Score:3, Interesting)
Does it run with more than 2 processors on the 390?
Is the 2 CPU limitation an X86-only thing that I'm ignorant of (quite possible)?
That's not to say that I don't love the BSD's, but they do have (or maybe they had) their limitations.
-Turkey
Re:Does this mean that... (Score:4, Informative)
I haven't been able to fully parse the boot log, so I'm not sure if it is utilizing more than one processor or not, but the thing that the write-up forgot to mention was that it was tested under the S/390 emulator (aka Hercules). I'm not sure what Hercules is and how it work per se... but it's still a huge jump forward for FreeBSD.
Re:Does this mean that... (Score:2)
Re:Does this mean that... (Score:2)
FreeBSD v5 will be a speed daemon.
Re:Does this mean that... (Score:2)
> I'm ignorant of (quite possible)?
I've ran NetBSD on a quad-Xeon machine.
- Hubert
Re:Does this mean that... (Score:1)
Cheers,
Jared
Re:Does this mean that... (Score:3, Informative)
The good news is that it works remarkably well on 1 or 2 cpu systems. It beat the performance of Linux 2.2 kernels and still gives 2.4 kernels a run for their money in most situations. When you start running mores cpus then performance will only go up a little bit so it really isn't worth it at that point.
FreeBSD 5.0 will not have this limitation and will scale nicely. I'm just not sure how far it will go at first but you can be sure that it will improve from there now that a decent setup for SMP is in place now (with 5.0).
Re:Does this mean that... (Score:2)
The good news is that it works remarkably well on 1 or 2 cpu systems. It beat the performance of Linux 2.2 kernels and still gives 2.4 kernels a run for their money in most situations. When you start running mores cpus then performance will only go up a little bit so it really isn't worth it at that point
Do you have any benchmarks or anything (non-anecdotal)? I've been looking for a side-to-side type comparison for awhile.
Re:Does this mean that... (Score:2, Interesting)
I've seen some inhouse testing with identical hardware with both Linux 2.4 and FreeBSD 4.x tuned for performance. They would go back and forth in some real-world testing with the edge in raw performance going to Linux more often. Usually the differences weren't that much but sometimes there was a clear winner.
I would suggest that you get some hardware to test with and use both on this hardware in the situation(s) that you are going to put it in. It is real easy to get a system that each can support the hardware. If one or the other doesn't support the hardware you want or very well then go with the OS that supports the hardware well.
For me, the real choice is outside of the scope of raw performance since they are fairly close (close enough for me at least). A couple factors I look at.
1. A moderate to heavy loaded FreeBSD box still responds well while a Linux box under the same load will become extremely sluggish or unresponsive.
2. Updating the system is more solid in FreeBSD than Linux. Apt-get and other mechanisms work well but usually have more issues for me at least. Following -stable is very easy in FreeBSD.
3. Experience. Which system is the admin more comfortable with? If they have experience BSD or other traditional Unix then FreeBSD is the way to go. If they are more comfortable with Linux then it really needs to factor in to the choice.
4. Number of processors. If the number goes over 2 then you need to go Linux or look into FreeBSD 5.0 if you can wait or run the development versions. Probably Linux will suit you better on your quad+ box for now at least. It will take some time to wring the most out of FreeBSD 5.x
These are all opinions and suggestions. Use both in your setup and see which works for you. Alternately you could tell them both to take a hike and get an Irix box. That would open a whole new set of things to look at.
Re:limits (Score:2)
Re:Does this mean that... (Score:1, Informative)
Re:I'll check it out as soon as... (Score:3)
The page for the port says it's being done on Hercules [conmicro.cx], which is a System/390 and z/Architecture simulator; the Hercules page claims it runs on Linux and 32-bit Windows, but it can probably be made to run on UNIXes other than Linux as well, assuming it doesn't Just Work out of the box.
Re:I'll check it out as soon as... (Score:2, Informative)
There is a FreeBSD port of the emulator in /usr/ports/emulators/hercules; if you're running another BSD flavor and it doesn't have this port you'll likely have an easy time importing it.
Re:I'll check it out as soon as... (Score:2)
As noted, it's in the FreeBSD packages tree. I plan to get it running on MacOS X by the time the next release is out, now that I've got a Mac to run it on. It will run on most Unixoid OSes with tweaking to remove Linux-specific SCSI tape and TUN/TAP code.
More the merrier (Score:2, Interesting)
The more platforms supported, the merrier it will be.
Although, I don't expect to see FreeBSD on anywhere near the same number of platforms as NetBSD.
I'd like to see FreeBSD 5 running on RS/6000 hardware... That would be nice
Re:More the merrier (Score:4, Informative)
Re:More the merrier (Score:2)
NetBSD isn't on that platform (Score:2, Interesting)
Re:NetBSD isn't on that platform (Score:2, Insightful)
At least, that's what the head guy said in his recent interview.
Re:NetBSD isn't on that platform (Score:2)
Server type things, of course!
- Hubert
Re:NetBSD isn't on that platform (Score:2)
Re:NetBSD isn't on that platform (Score:2, Funny)
The thing I think is interesting about this port is that it puts freebsd on the s/390 while NetBSD isn't.
Think about it more practically:
A lot more people have toasters than have S/390's.
Re:Kind of silly. (Score:2, Insightful)
Um, it's just one guy doing this, so far and he'll do whatever interests him, obviously. If all free software people did what was "needed" and not what was personally interesting to them, commercial OS's would be extinct by now.
Re:Kind of silly. (Score:3, Insightful)
Indeed. Look at all the poor free software that there are multiple, independent, poor versions of. (e.g. DVD/mpeg players, web browsers, word processors, financial software, file managers). If people had coordinated, with a goal of producing what was needed, instead of each writing their own software independently because of what they personally wanted, the free solutions would be so vastly superior to the commercial solutions that only the free ones would survive. Unfortunately, this almost never happens.
Re:Kind of silly. (Score:1)
It's doubtful, though. IBM has officially blessed Linux for mainframes, and it's unlikely that any organization would give up the support they can get for that to run *BSD, regardless of the merits.
Garg
I wonder what it would be like ... (Score:1)
to have a Beowolf cluster -- oh never mind!
Before people get too excited about topics such as SMP, kernel threads, and I/O devices -- it only partially boots on a mainframe emulator. This is a VERY LONG WAYS OFF from asking, "so where can I download the ISO images?".
Re:I wonder what it would be like ... (Score:3, Interesting)
Yeah...the project isn't to the point where it's usable for anything but system hacking yet.
on a mainframe emulator.
If it runs on Hercules, it'll run on the real hardware. Before you pooh-pooh the use of an emulator, consider that Alan Cox uses Hercules for S/390 work (not all of it, but quite a bit).
Re:*BSD is dying (Score:1)
Re:*BSD is dying (Score:1)
A close aquaintence of mine WROTE
386BSD!
Actually, if people wern't such idiots,
386BSD would be popular!
But, I must admit, other *BSD's than
386 are very crappy.
P.S. Do not insult 386BSD! Its the other
*BSD's that are bad!
Re:*BSD is dying (Score:1)
Terminals (Score:1)
Unix machines have had more processing power than S/390s for years now, but they did commonly not support block-mode terminals (like S/390's 3270), and therefore it was difficult to implement host-/terminal-based applications on unix-systems without causing too much network traffic.
AS/400's (aka iSeries 400) have got 5250-terminals, which are quite the same as the 3270's, but a few years ago AS/400's did not have as much processing power and/or disk capacity as mainframes. Recent AS/400's are mainly based on pSeries-like hardware, with the same powerful POWER4 processors, lots of cache memory and lots of main storage (RAM).
Since the biggest AS/400s are now more powerful than the biggest z/900 but also are more mainframe-like than the pSeries, it could be possible, that IBM is going to replace the S/390's with the more modern AS/400's.
don't hate *BSD (Score:1)