Slashdot Mirror
IBM Supporting Linux On Power Processors
aheath writes "IBM issued a press release today titled 'Businesses Embrace Linux and IBM POWER Technology', indicating that: 'IBM eServer pSeries and iSeries systems are available for SUSE and Red Hat Linux operating systems (OS) on IBM's POWER microprocessor-based architecture.' CNET News are running a story about this, too, explaining: 'IBM has put more weight behind its effort to attract customers to Linux that runs on its own Power processors, an initiative that distinguishes Big Blue from its competitors in the server market.' IBM has also signed up 300 vendors to provide software to run under Linux on Power processors."
Gee... they're already working on it.
> As Linux gets more and more momentum behind it, it starts to seem like it will be the last OS that will ever make serious inroads into the general public.
You may be right, but I don't see how that conclusion could possibly follow from that observation. Linux may be the next OS to make serious inroads, but how could we possibly expect it to be the last one that does?
Sheesh, evil *and* a jerk. -- Jade
Yellow Dog for the G5 has been available for a while; it's working OK in my office.
I looks like binary only centrino support is on its way with an open source solution coming later, according to news.com
"DRM is like violence: if it doesn't work, use more."
Well, first thing, of course, is that Apple CPUs are IBM CPUs. For the G5, the whole design is from IBM; but even the G4s are fabbed by IBM (though designed by Motorola). So big blue likes the PowerPCs all around.
Still, the IBM announcement makes some sense of my IBM developerWorks editor's push for a rush job on an article about Linux on PowerPCs. Despite the rush, I think it turned out well. Take a look at:
Buy Text Processing in Python
... Intel and AMD lead, hell, even Apple's CPUs are more popular than IBM's.
More popular by whose standard? Their market share may be negligle but it is definitely high-end. Companies (banking, FAA, etc.) will pay top dollar for machines that simply don't fail.
I work for an orgainzation that requires a minimum of 99.9999% uptime. We have been using IBM RISC-based gear since the RT came out because of this. We used some Sun gear for a while but those just didn't cut the mustard.
Arf!
It's interesting you say the Mac is a joke for serious number crunching, and then immediately mention bioinformatics. That's actually an area where Apple has had some success. The Mac has always been fairly popular in biology for some reason, and a lot of the requests which convinced Apple to make a rack-mount Mac came from that sector. Apple is now offering pre-built Xserve clusters loaded up with G5-optimized bioinformatics software. See Apple Workgroup Cluster for Bioinformatics.
Of course, it's IBM's chip at the heart of this.
This space unintentionally left unblank.
With the exception of the PowerPC 601 and the possible exception of the G5 (I don't know about this one), the POWER processors have more instructions than their PowerPC counterparts of the same generation and based on the same core. This is why MPC601 is supposedly so great clock for clock, but also so expensive and slow. In POWER, efficiency and reliability are the primary selling points, as you say. PowerPC's are speed and low power consumption (Well, low power consumption USED to be a selling point.) So, PowerPC binaries might run fine on POWER, but in general, the converse is not true.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Besides, ever worked on labs that _really_ need to do serious number crunching (hint, hint)? They swear by IBM. PC clusters are a joke, Macs even more so.
What? Do you think that NCSA doesn't do "serious number crinching" with THIS ? Guess what? They're Dells, powered by Intel Silicon.
How about Lawrence Livermore National Laboratory it's running a linux cluster with 1,116 nodes.
One of his core business models for furthering the bio-informatics idea was to contact IBM and get them to design CPUs that would optimize the algorithms for certain vector and matrix operations. And he had enough information from IBM contacts who confirmed that if the idea proved viable, they would do so.
Next are you going to tell us that this guy works in the "Double '0'" section of the British Secret Service?
IBM has made some revolutionary discoveries in microprocessor design. IBM is still one of the most innovative companies out there. IBM still produces some serious Big Iron, but let's not over state it.
LK
"Hi. This is my friend, Jack Shit, and you don't know him." - Lord Kano
IBM has an enormous archive of patents, many of which are so broad that they're used in Linux (in some way or another). If they were to release their own Linux distro, they'd effectively be releasing their 'IP' under the GPL, and complicate the situation.
That's the reason I've heard before, anyway!
While I agree and *don't* call Linux GNU/Linux, there are operating systems that execute Linux binaries without modification and they are not Linux. The BSD Linux compatability layer comes to mind.
A firewall can not protect you from yourself. Turn off what you do not need. Do not use the firewall to do your work.
The same is IMO inevitable with Linux because it has solved the main issue of how to provide the necessary services on any platform in a non-partisan manner. ...
I do remember the old protocols as I ran a test lab with a dozen servers -- mostly using different propriatory protocols.
Linux isn't based on propriatory interfaces, so making a "Linux clone" work transparantly with Linux binaries -- let alone source code -- is not only possible, it has been done already -- even for the PPC!
A firewall can not protect you from yourself. Turn off what you do not need. Do not use the firewall to do your work.
Nope, it's open first and the cost is a side-effect. A very tempting side-effect, but not the point. Otherwise, freeware should be gaining in popularity and wide spread use.
A firewall can not protect you from yourself. Turn off what you do not need. Do not use the firewall to do your work.
A Debian install, on top of a NetBSD kernel. The shell is bash (GNU), and this interprets all of the init scripts. The loader comes from the GNU project. The next thing to be launched is XFree86 (not GNU, and not GPL'd). On top of X, runs gdm (GNU again), which provides a graphical login screen to the user. Once you log in, you are presented with the GNOME desktop environment (no prizes for guessing what the G in GNOME stands for...) You browse the web a bit with Epiphany (can you guess?) and mangle some figures with GNUmeric. Perhaps you decide you don't like Epiphany, and download the Linux version of Opera. Of course, this runs fine on NetBSD's Linux binary compatibility layer.
By your definition, this is a Linux system (i.e. it executes a Linux ELF binary). I would contend, however, that there is a lot more GNU (i.e. lots) than there is Linux (i.e. none) in this system.
I am TheRaven on Soylent News
My first Unix was v7, 25 years ago.
You still move files, dd things, and have a shell script, make, cc, and so on.
Yeah, a few options have changed here and there, but the fundimentals of the OS are easily familar if you time travel fowards or backwards a couple of decades.
It brings a lot of it... But not all, by any means.
Many programs which manipulate binary data are hard-coded with the assumption that the data order is little-endian, as it is on Intel. These programs will break on a big-endian architecture like Power.
Also, some programs may rely on memory protection functionality which is not available on a particular architecture (for example, on Intel, it's impossible to mark a page as "execute-only." Programs which depend on being able to mark pages as "execute-only" thus cannot function on x86 processors.)
As another example, a multimedia application which relies on certain real-time assumptions may not function on a platform which cannot make those real-time guarantees (perhaps because it has a crappy interrupt architecture).
And of course, there's always software that's mainly written in C, but also has little portions written in assembler. Clearly, these programs won't just magically compile on a completely different processor.