Redhat is probably the best thing that's happened to Linux. Nobody will ever following anything without a strong backer, and Redhat is it. With SGI hopping on board and the government's adoption of Linux that will follow, things will even get better. But it's Redhat that got the ball rolling, and Redhat that employs a lot of the full time kernel developers.. And that's good for all of us. Three cheers!
From a company that's tried to do firewire stuff..
on
Is firewire dying?
·
· Score: 1
I'm told by guys at our company actually working on FireWire boards that the standard is muddled, messed up, and the royalties to Apple make it a real pain in the neck to do anything with it. Intel is in favor of open standards, like PCI, not proprietary standards. So, unless things change, FireWire will probably remain rather useless unless you have a Mac.
An educated opinion is a well-informed opinion. The following is my list, in order of decreasing complexity and (what I think) are admirable design characteristics and flaws:
1. IRIX - fully multithreaded, very scalable, but VERY expensive (prohibitively in most app areas) 2. WinNT - fully multithreaded, flexible, not too scalable, though it has a large learning curve and suffers from backward compatability 3. Tru64 (formerly Digital UNIX) - microkernel based, flexible, fast, scalable, great networking OS 4. QNX - microkernel, RTOS, small footprint. Main drawbacks are expense, copying overhead in the OS and only (currently) runs on X86 5. BeOS - microkernel, fast, efficient, great development tools, portable. Doesn't scale to really big boxes.. At least not yet. 6. Solaris - scalable, though it's too close to classic UNIX architecture for my taste and ships rather broken until you apply a gazillion patches. 7. Linux - older architecture kernel, over-hyped, doesn't scale well for the lack of fine grained kernel locks, not fully POSIX compliant (yet). Is easy to develop on and has a wide base of portability; it's also cheap and supports a lot of lower end hardware, which makes it great for clustering (MIMD supercomputers) and research. 8. FreeBSD - fast, efficient, but supports only X86 and doesn't scale very well. With Alpha support, it's winning some of my favor. 9. VxWorks - great for embedded stuff, not much use to the average consumer; it's a RTOS. 10. LynxOS - another RTOS, POSIX compliant, runs on PPC and X86, but it's rather broken unless you're reserving it for a dedicated application. 11. MacOS - it's pretty but it's reminiscent of Win 3.1. MacOS X will change that.. But for now the Apple OS is rather worthless IMHO. 12. HPUX - it's a kludge; it's a miracle it works. HP makes good printers. That's about all I give them credit for. 13. Win95 - it's not an OS, get over it.
VMS would be close to the top of my list, but it's dead.. Anyway WinNT is largely based on the good things of VMS from a computer science perspective, as is Tru64; hence they're on the top of my list.
This opinion is based on ease of use, flexibility, scalability, TCO, portability, and of course- adherence to classical OS theory of each OS. My two cents.
Thank you. About time somebody said it. It's the *IGNORANT FOOL* users that spread their misinformed opinions about Linux that actually hurt it. Rather than *acknowledging its weaknesses* (HINT - even LINUS does - read the kernel mailing list) - those zealots are the ones who only hurt Linux as serious users point to them and say "ha, I'll be damned if I'll run an OS if people can only spout off such garbage."
I used to say the same thing when I ran OS/2. I've used them *all*, and I know the advantages and disadvantages as well as the architectures of each. Now if only everyone else did and could remain civilized...
2048. Go read the latest supercomputer list. A single image IRIX-based Origin 2000 supercomputer is number six on the list. But the claim about IRIX being the first 64-bit OS is true if you are talking a *single image* system.
Of course not. But they are transitioning their *workstations* to IA-64.
And as far as IRIX goes, they have a good reason NOT to port it to IA-64. It is much too reliant on the underlying hardware for primitives, and likely would not perform as well on a large IA-64 box (that is assuming that IA-64 will scale to 64+ CPUs).
>Not to mention that these are real companies that are "jumping on a bandwagon." This isn't some kid that decides to install linux so he can impress the hackers in the next grade up. SGI is a real company, with a shitload of money, and ok, maybe they're in a little trouble, but they're not going to "jump on a bandwagon" just for the thrill ride, or to impress some people. They're doing it because they think they'll make money off of it.
Not to mention that if SGI went away tomorrow, they'd leave a BIG vacuum that no one else would be able to fill. They won't be going away anytime soon, and anything they contribute will be worth tenfold what will come out of the Linux community for free... That's because IRIX was developed with a combination of both a "get it right" and "brute force" approach that is the attitude necessary to develop a cutting edge OS that works. Linux developers don't have the time or resources to do that.. At least not until now. I'm looking forward to seeing where it will go from here! Surely, the sky is now the limit!
You guys have a LOT to learn about big machines if you think Linux will ever run on a 2048 processor box that's used for what the Cray is used for. Sheesh:(
Ted T'so is working on ext3. The jfs extensions in 2.4 will be add-ons to ext2 based on reiserfs.
Re:Yeah, but $8000 for an entry-level box???
on
SGI's Linux Server
·
· Score: 1
Also, why the hell dont they make a move to the EV6 with the k-7's... that will surely rock the house! Read some pages on their site. They're migrating toward IA-64 and ditching MIPS within the next couple of years. This is one step in their migration. More than likely, they'll release one last round of Onyx2's and Origin 2000's based on the R12k CPU, and then we'll start seeing all new BIG machines. IA-32 is just their way of testing the waters.
Unless you're 100% I/O bound on DMA controllers, I guarantee that NT and IRIX both will kick Linux in the *** on 4+ CPUs, since both have 100% reentrant multithreaded kernels, while the Linux kernel is not reentrant. This is changing, but for now we have to live with it on big boxes. (see the post Re: David Miller; I know he's done a lot of testing on 8+ CPUs and the scaling is quite horrible; Solaris SMP really sucks and it outperforms Linux about 2:1 on 14 CPUs)
I'll bet that they've been doing some tinkering with the Linux kernel. I've noticed quite a few posts on the Linux kernel mailing list lately from IRIX programmers inside SGI.. I thought it was kinda interesting... SGI is a great company, and having them getting involved in Linux is like having a major league ballplayer join your softball team.:)
Give them a week to resolve it. If they still won't, find yourself a non porno-hosting ISP and you won't have to worry about it. Above all, let them know what they did by not telling you about the ban was wrong and unfair to you and that if they don't resolve it, they'll lose more customers. pair.com is always cheap, fast, and open. If they don't keep their users happy, someone else will. The Internet is a big place, after all.
Slashdot Mirror
Redhat is probably the best thing that's happened to Linux. Nobody will ever following anything without a strong backer, and Redhat is it. With SGI hopping on board and the government's adoption of Linux that will follow, things will even get better. But it's Redhat that got the ball rolling, and Redhat that employs a lot of the full time kernel developers.. And that's good for all of us. Three cheers!
I'm told by guys at our company actually working on FireWire boards that the standard is muddled, messed up, and the royalties to Apple make it a real pain in the neck to do anything with it. Intel is in favor of open standards, like PCI, not proprietary standards. So, unless things change, FireWire will probably remain rather useless unless you have a Mac.
An educated opinion is a well-informed opinion. The following is my list, in order of decreasing complexity and (what I think) are admirable design characteristics and flaws:
1. IRIX - fully multithreaded, very scalable, but VERY expensive (prohibitively in most app areas)
2. WinNT - fully multithreaded, flexible, not too scalable, though it has a large learning curve and suffers from backward compatability
3. Tru64 (formerly Digital UNIX) - microkernel based, flexible, fast, scalable, great networking OS
4. QNX - microkernel, RTOS, small footprint. Main drawbacks are expense, copying overhead in the OS and only (currently) runs on X86
5. BeOS - microkernel, fast, efficient, great development tools, portable. Doesn't scale to really big boxes.. At least not yet.
6. Solaris - scalable, though it's too close to classic UNIX architecture for my taste and ships rather broken until you apply a gazillion patches.
7. Linux - older architecture kernel, over-hyped, doesn't scale well for the lack of fine grained kernel locks, not fully POSIX compliant (yet). Is easy to develop on and has a wide base of portability; it's also cheap and supports a lot of lower end hardware, which makes it great for clustering (MIMD supercomputers) and research.
8. FreeBSD - fast, efficient, but supports only X86 and doesn't scale very well. With Alpha support, it's winning some of my favor.
9. VxWorks - great for embedded stuff, not much use to the average consumer; it's a RTOS.
10. LynxOS - another RTOS, POSIX compliant, runs on PPC and X86, but it's rather broken unless you're reserving it for a dedicated application.
11. MacOS - it's pretty but it's reminiscent of Win 3.1. MacOS X will change that.. But for now the Apple OS is rather worthless IMHO.
12. HPUX - it's a kludge; it's a miracle it works. HP makes good printers. That's about all I give them credit for.
13. Win95 - it's not an OS, get over it.
VMS would be close to the top of my list, but it's dead.. Anyway WinNT is largely based on the good things of VMS from a computer science perspective, as is Tru64; hence they're on the top of my list.
This opinion is based on ease of use, flexibility, scalability, TCO, portability, and of course- adherence to classical OS theory of each OS. My two cents.
You don't even know what SIMD, MIMD and MISD mean, do you?
Thank you. About time somebody said it. It's the *IGNORANT FOOL* users that spread their misinformed opinions about Linux that actually hurt it. Rather than *acknowledging its weaknesses* (HINT - even LINUS does - read the kernel mailing list) - those zealots are the ones who only hurt Linux as serious users point to them and say "ha, I'll be damned if I'll run an OS if people can only spout off such garbage."
I used to say the same thing when I ran OS/2. I've used them *all*, and I know the advantages and disadvantages as well as the architectures of each. Now if only everyone else did and could remain civilized...
..it scales up to 256 processors
2048. Go read the latest supercomputer list. A single image IRIX-based Origin 2000 supercomputer is number six on the list. But the claim about IRIX being the first 64-bit OS is true if you are talking a *single image* system.
Go get yourself a kernel book or two and READ about why Linux won't scale well. Then let us know what you learn.
Of course not. But they are transitioning their *workstations* to IA-64.
And as far as IRIX goes, they have a good reason NOT to port it to IA-64. It is much too reliant on the underlying hardware for primitives, and likely would not perform as well on a large IA-64 box (that is assuming that IA-64 will scale to 64+ CPUs).
The truth is out there... Now go find it!!
>Not to mention that these are real companies that are "jumping on a bandwagon." This isn't some kid that decides to install linux so he can impress the hackers in the next grade up. SGI is a real company, with a shitload of money, and ok, maybe they're in a little trouble, but they're not going to "jump on a bandwagon" just for the thrill ride, or to impress some people. They're doing it because they think they'll make money off of it.
Not to mention that if SGI went away tomorrow, they'd leave a BIG vacuum that no one else would be able to fill. They won't be going away anytime soon, and anything they contribute will be worth tenfold what will come out of the Linux community for free... That's because IRIX was developed with a combination of both a "get it right" and "brute force" approach that is the attitude necessary to develop a cutting edge OS that works. Linux developers don't have the time or resources to do that.. At least not until now.
I'm looking forward to seeing where it will go from here! Surely, the sky is now the limit!
..Only if IRIX would.
You guys have a LOT to learn about big machines if you think Linux will ever run on a 2048 processor box that's used for what the Cray is used for. Sheesh:(
>Of course it seems SGI can't make up it's mind
>for any great period of time anymore,
Perhaps you have not read their business plan?
Take a look. This is yet one step of a long, thoughtfully laid out path.
The information is there if you choose not to remain ignorant..
Ted T'so is working on ext3. The jfs extensions in 2.4 will be add-ons to ext2 based on reiserfs.
Also, why the hell dont they make a move to the
EV6 with the k-7's... that will surely rock the house!
Read some pages on their site. They're migrating toward IA-64 and ditching MIPS within the next couple of years. This is one step in their migration. More than likely, they'll release one last round of Onyx2's and Origin 2000's based on the R12k CPU, and then we'll start seeing all new BIG machines. IA-32 is just their way of testing the waters.
Unless you're 100% I/O bound on DMA controllers, I guarantee that NT and IRIX both will kick Linux in the *** on 4+ CPUs, since both have 100% reentrant multithreaded kernels, while the Linux kernel is not reentrant. This is changing, but for now we have to live with it on big boxes. (see the post Re: David Miller; I know he's done a lot of testing on 8+ CPUs and the scaling is quite horrible; Solaris SMP really sucks and it outperforms Linux about 2:1 on 14 CPUs)
I'll bet that they've been doing some tinkering with the Linux kernel. I've noticed quite a few posts on the Linux kernel mailing list lately from IRIX programmers inside SGI.. I thought it was kinda interesting... SGI is a great company, and having them getting involved in Linux is like having a major league ballplayer join your softball team. :)
Give them a week to resolve it. If they still won't, find yourself a non porno-hosting ISP and you won't have to worry about it. Above all, let them know what they did by not telling you about the ban was wrong and unfair to you and that if they don't resolve it, they'll lose more customers. pair.com is always cheap, fast, and open. If they don't keep their users happy, someone else will. The Internet is a big place, after all.