SGI: Discussion

Linus about SGI and Linux - Page 1

http://www.h-online.com/open/features/Interview-Linus-Torvalds-I-don-t-read-code-any-more-1748462.html

Quote:
SGI in particular worked a lot on scaling past a few hundred CPUs. Their initial patches could just not be merged. There was no way we could take the work they did and use it on a regular PC because they added all this infrastructure to work on thousands of CPUs. That was way too expensive to do when you had only a couple.

I was afraid for the longest time that we would have the high-performance kernel for the big machines, and the source code would be separate from the normal kernel. People worked a lot on just making sure that we had a clean code base where you can say at compile time that, hey, I want the kernel that works for 4,000 CPUs, and it generates the code for that, and at the same time, if you say no, I want the kernel that works on 2 CPUs, the same source code compiles.

It was something that in retrospect is really important because it actually made the source code much better. All the effort that SGI and others spent on unifying the source code, actually a lot of it was clean-up – this doesn't work for a hundred CPUs, so we need to clean it up so that it works. And it actually made the kernel more maintainable. Now on the desktop, 8 and 16 CPUs are almost common; it used to be that we had trouble scaling to 8, now it's like child's play.

_________________
:Octane2: 2xR12000 400MHz, 4GB RAM, V12 GFX
SGI - the legend will never die!!
That's why I was so surprised when SCO went after IBM, since SGI was responsible for basically all of the non-trivial SMP and NUMA code in Linux. IBM's contributions were relatively paltry by comparison...

_________________
Choosing stones, big enough to drag me down...
The whole suit seemed like a "hail Mary" attempt by SCO to bring revenue. So it did not make sense for SCO to waste time suing a company with no cash (I think SGI was already circling the drain by the time this nonsense was brought).

Of course, given the flimsy case SCO had, one has to question the wisdom of patent trolling IBM, a company with a track record of not being afraid of using aggressively their rather sizable patent and lawyer portfolio.

Anyhow, the NUMA stuff probably postdates the Unix codebase which SGI may have licensed at some point.

_________________
"Was it a dream where you see yourself standing in sort of sun-god robes on a
pyramid with thousand naked women screaming and throwing little pickles at you?"
R-ten-K wrote:
Of course, given the flimsy case SCO had, one has to question the wisdom of patent trolling IBM, a company with a track record of not being afraid of using aggressively their rather sizable patent and lawyer portfolio.

The magic word is "kickback".

"Management team" at SCO knows it isn't long for this world but there's still money coming in. Pay the creditors ? Why should we do that ? We can pay Cousin Ernie to sue IBM instead ... and later on we might need a job. Cousin Ernie has lots of friends.

There's no need for a conspiracy when they all think the same.

Subject of this thread is kinda funny though. I totally lost interest in Linux when Mr Torvalds stated unequivocably that Linux would not ever ever ever be smp-capable. "This is not a corporate operating system."

yeah well ......
hamei wrote:
"This is not a corporate operating system."

yeah well ......


How'd that work out? Tell me he's not getting any kind of kickbacks

_________________
-ks

:Onyx: :Onyx: :Crimson: :O2000: :Fuel: :Octane: :Octane2: :PI: :Indigo: :Indigo: :O2: :O2: :Indigo2: :Indigo2: :Indigo2IMP: :Indigo2IMP: :0300: :Indy: :Indy: :320: :540: :1600SW: :1600SW: :hpserv:

See them all >here<
Are there any computers in the Top 500 that aren't running Linux? Maybe a few but not many... http://www.top500.org/

It's so funny in TFA where it says Linus' job at Transmeta `didn't work out,` well sure the company tanked but he and all the other founders made out like bandits off the stock...

_________________
Choosing stones, big enough to drag me down...
Commercial Unix used to have the server and workstation market, BSD hadn't gained much traction outside academia, and Linux was almost exclusively for home users running PC's. For a long time, SGI and other workstations were viewed as exotic and expensive machines that few people had access to. These days, two processors doesn't seem that impressive since everyone has dual and quad core systems, but at one point, the situation was very different.

http://news.cnet.com/SGI-to-slash-workstation-prices/2100-1001_3-213534.html
Quote:
The Octane line's entry-level product, which comes with a 225-MHz R10000 MIPS processor, 128MB of memory, a 4GB hard drive, and a 20-inch monitor, will fall to $17,995 from $19,995. The pricing action comes two months after the company introduced it. An Octane system featuring 250-MHz R10000 processor, meanwhile, will drop from $38,995 to $24,995.

I wonder how many people had such a workstation in their homes in 1998?

Like the dwarves in Lord of the Rings, the big Unix players delved too greedily and too deep.

_________________
Debian GNU/Linux on a ThinkPad, running a simple setup with Fvwm.
vishnu wrote:
well sure the company tanked but he and all the other founders made out like bandits off the stock...


Torvalds was not a founder of Transmeta, and he left well before the company folded. I don't know how one can "make out like a bandit" off a tanked stock exactly.

_________________
"Was it a dream where you see yourself standing in sort of sun-god robes on a
pyramid with thousand naked women screaming and throwing little pickles at you?"
Quote:
I wonder how many people had such a workstation in their homes in 1998?


me! (indigo 2@250Mhz SI)

_________________
:Onyx2:
R-ten-K wrote:
I don't know how one can "make out like a bandit" off a tanked stock exactly.

One way is to short it ... I have an acquaintance who's a fraudmeister. He was doing the same "Booming China market ! get in at the bottom while you still can !" schtick like everybody else but the field was too crowded. He's the exception, an MBA with an IQ over ten. So he invented a "debunk the China stock" game. He finds a likely company that everyone else is peddling and does real on-the-ground research.

For example, here's a company doing two million a month, sales increasing 30% a year, has the ear of the local government, privately owned, nimble, versatile, modern, a lean mean killing machine woohoo yippee, going to be huge, every buzzword that the moron American "investors" drink with their dinner.

The trouble is, all ( I mean all ) these places are frauds. So Carsten goes there, hangs out by the front gate of the shabby little 3,000 square foot building out in some desolate village somewhere, takes photos of the five employees arriving each day and the one delivery a week. He shorts the high-flying "investment" then releases a true report on the company. Stock crashes, he makes out like a bandit on the tanked stock. You can look it up, I think his company name is Muddy Waters.

Repeat, repeat. There's a large supply of idiot investors. He's careful to not give the impression that all Chinese companies are frauds because that might chase away the fools but in fact, all of them that have stock are. As are all the American companies who issue stock. It's a con game. The entire United States has been fleeced*. In the real world the finance sector contributes absolutely nothing - but they sure do rake in the cash.

Really, I have to admire him. Between the never-ending supply of drooling American investors and fraudulent Chinese companies, he can have a career that will last a long time. And he's doing good while he's at it.

* e.g., Hewlett-Packard. Ten billion dollars and 60,000 employees down the toilet. Where is Mark Hurd ? duhn, I dunno. Gone. Where is Carly Fiorina ? dunh, I dunno. Gone, I guess. Gee, sorry, we did the best we could ... they went on the news and said they took responsibility, what more do you want ?
R-ten-K wrote:
I don't know how one can "make out like a bandit" off a tanked stock exactly.
On the day of Transmeta's IPO Linus' shares were worth, at one point, $3,360,000. Now I'm sure he didn't sell that many shares at that price, ever, but he still sold a lot, and made a lot of money doing it, as did everyone else who got in on Transmeta's ground floor. Sure Transmeta tanked in the end, but they flew pretty high for a while...

_________________
Choosing stones, big enough to drag me down...
Linux is the basis of an ecosystem that generates hundreds of millions (if not billions) of $. The guy who created and maintains the kernel which makes everything tick may have raked in a couple of million $ in the process. Apparently that is wrong, how exactly?

_________________
"Was it a dream where you see yourself standing in sort of sun-god robes on a
pyramid with thousand naked women screaming and throwing little pickles at you?"
Linux and gnu; and there's your answer, it's the gpl. If you want to make vast fortunes, keep your source closed and hope for the best, if you want your code to be used by potentially the most amount of people, open your source. And that points out a big difference between the gpl and bsd licenses, none of the bsd developers are living in houses as nice as Linus'...

_________________
Choosing stones, big enough to drag me down...
R-ten-K wrote:
Linux is the basis of an ecosystem that generates hundreds of millions (if not billions) of $. The guy who created and maintains the kernel which makes everything tick may have raked in a couple of million $ in the process. Apparently that is wrong, how exactly?

It's wrong because it's based on a lie. The whole point to Linux was "the evil corporations won't let us have Unix at a decent price so we'll make our own. This is for the commmmuuuunity ." 90% of the work was done by others for free, based on these claims. I seriously doubt that any of those people would have lifted a finger if the point of the exercise was to create another cash cow for IBM and Linus Torvalds.

I don't hate the man or anything but if you're going to be honest, his entire career and stature is based on a lie. Stallman may be a goofball but at least he walks the walk. Torvalds is a hypocrite.

Okay article though - you have to laugh :

Linus Torvalds wrote:
It used to be way flatter. I don't know when the change happened, but it used to be me and maybe 50 developers -- it was not a deep hierarchy of people. These days, patches that reach me sometimes go through four levels of people. We do releases every three months; in every release we have like 1,000 people involved. And 500 of the 1,000 people basically send in a single line change for something really trivial -- that's how some people work, and some of them never do anything else, and that's fine. But when you have a thousand people involved you can't have me just taking patches from everybody individually.

We're doing really well. The kind of pain points we had ten years ago just don't exist any more. And that's largely because we used to be this flat hierarchy, and we just fixed our tools, we fixed our work flows. And it's not just me, it's across the whole kernel there's no single person who's in the way of any particular workflow.

Cathedral

Bazaar

ha
ha
ha

:P
mia wrote:
Quote:
I wonder how many people had such a workstation in their homes in 1998?


me! (indigo 2@250Mhz SI)

:o Lucky guy!
hamei wrote:
It's wrong because it's based on a lie. The whole point to Linux was "the evil corporations won't let us have Unix at a decent price so we'll make our own. This is for the commmmuuuunity." 90% of the work was done by others for free, based on these claims. I seriously doubt that any of those people would have lifted a finger if the point of the exercise was to create another cash cow for IBM and Linus Torvalds.

I don't hate the man or anything but if you're going to be honest, his entire career and stature is based on a lie. Stallman may be a goofball but at least he walks the walk. Torvalds is a hypocrite.

By the time the Linux kernel was developed, GNU tools had already been popular on commercial Unixes for quite some time. For example, if someone wanted a "grep" that was faster and without arbitrary limits, they could use the GNU version. That was their selling point -- more options, fewer limitations, portable, and.... emacs. Then in the early 90's, Linus came along and wrote a kernel for his 386, compiled the GNU userland along with it, put in some glue, and then everyone started calling the whole system "Linux." The interesting thing is that GNU code far outweighs kernel code when looking at the whole system. Without the GNU project, "Linux" would have to import all the tools from BSD to even be usable.

Later, Stallman contacted Linus about working together toward a common goal of "software freedom," but of course Torvalds didn't really care about that stuff, and he still doesn't to this day. He just considers the GPL to be a tool that helps him keep any derivative code in the official kernel. There has always been this conflict between the "free software" movement and others who don't care at all about those principles. Eric Raymond is another person who is like this. Coining the term "open source" was all about weakening the stance on free software, and sweeping that set of ideas under the carpet.

The Linux kernel is not so indispensable, though. Considering the project "Debian GNU/kFreeBSD", it shows that a distribution can even change kernels while keeping almost all the other software the same. They could probably also even create a "Debian GNU/Illumian" or a "Debian GNU/Minix3" if they wanted to.

_________________
Debian GNU/Linux on a ThinkPad, running a simple setup with Fvwm.
The commercial Unix vendors tried to do it all themselves, and by the time they realized they couldn't compete with the specialists (Intel, Microsoft, Nvidia, Linux, GNU), it was too late. Jim Clark was ringing the warning bell as early as 1993, but he was a voice crying out in the wilderness. So he stormed out of SGI in a blaze of smoking carpet fibers and made a fortune at Netscape, which, just like Transmeta, was a company that only made money for it's pre-IPO stockholders...

_________________
Choosing stones, big enough to drag me down...
vishnu wrote:
Linux and gnu; and there's your answer, it's the gpl. If you want to make vast fortunes, keep your source closed and hope for the best, if you want your code to be used by potentially the most amount of people, open your source. And that points out a big difference between the gpl and bsd licenses, none of the bsd developers are living in houses as nice as Linus'...


You can go to kernel.org, and download the Linux kernel for free. Furthermore, nothing in the GPL says that you can't charge for a product, or that you can't earn a living out of open source, only that you have to distribute the code and if someone modifies said code they have to contribute those modifications back. Here's what the gnu guys themselves have to say:
Quote:
Many people believe that the spirit of the GNU Project is that you should not charge money for distributing copies of software, or that you should charge as little as possible — just enough to cover the cost. This is a misunderstanding.

Actually, we encourage people who redistribute free software to charge as much as they wish or can. If this seems surprising to you, please read on.

The word “free” has two legitimate general meanings; it can refer either to freedom or to price. When we speak of “free software”, we're talking about freedom, not price. (Think of “free speech”, not “free beer”.) Specifically, it means that a user is free to run the program, change the program, and redistribute the program with or without changes.

http://www.gnu.org/philosophy/selling.html

_________________
"Was it a dream where you see yourself standing in sort of sun-god robes on a
pyramid with thousand naked women screaming and throwing little pickles at you?"
Then again Apple is probably making more money from a free software product (FreeBSD) than anyone else is making selling Linux-based systems, and it's precisely the GPL that precluded Apple from using Linux in OSX...

_________________
Choosing stones, big enough to drag me down...
It's a bit more complicated than that. At some point Apple actually developed their own version of Linux; mkLinux. However, Apple decided to purchase NeXT in the late 90s in order to use their OS/Software technologies to replace MacOS OS X is based on NextStep (or OpenStep or whatever it was named by then), which was an implementation of Mach using a BSD server/personality/userland and predates Linux actually. OSX is not FreeBSD, it uses parts/subsystems of FreeBSD (and NetBSD as well) for some of the stuff running on top of XNU (or whatever Apple calls their kernel) in order to keep the same BSD-on-top-of-Mach architecture inherited from NeXT.


Apple (or NeXT) never released as open source some of the run-time libraries however, since those were never derived from GNU sources. But Apple has use(d) the GNU toolchain extensively, gcc being the compiler of the NeXT/OpenStep/OSX/iOS ecosystem up to a few months ago. This is, NeXT and now apple have used (and included with their products) GNU-licensed software for a couple of decades at least. There is nothing about GNU that precludes it being used for commercial products.

I have no idea what this has to do with Torvalds's supposed moral shortcomings though.

_________________
"Was it a dream where you see yourself standing in sort of sun-god robes on a
pyramid with thousand naked women screaming and throwing little pickles at you?"
vishnu wrote:
Then again Apple is probably making more money from a free software product (FreeBSD) than anyone else is making selling Linux-based systems, and it's precisely the GPL that precluded Apple from using Linux in OSX...

I don't think so. XNU is opensource . I think this choice was made because Mac OS X evolved from NeXTSTEP and it wouldn't make much sense to replace the whole kernel with Linux.

_________________
:Indigo: :Indigo: :Indy: :Indy: :Indigo2: :Indigo2IMP: :Octane: :Fuel: Image