Alpha has had this probably from the beginning (at least since '94, when I saw this in an Alpha ISA document). I also recall seeing prefetch instructions on at least one VLIW design. Do (some) L2 caches enable the CPU to issue line/block prefetch commands?
>> If you don't need MS Windows to run MS office, then you don't need MS Visual Studio to write apps with...
Now _here's_ a terrifying thought - what if porting Office necessitates porting a part of MFC => Unix cross development on Visual Studio. Then you get Office for Linux, IE for Linux etc. _and_ a development environment from MS (even if they only use that in-house). This means that MS put their browser into Linux/Unix-space; now here's some serious competition for any browser on Unix (including Mozilla for sure). I wonder what kind of havoc that will wreak... I for one am looking forward to having Office on Unix - even if the product sucks as much as it does now, it will be common ground for replacement of Windows boxes with Unix. Then we have the possibility of migrating from MS Office to Star Office or whatever, since they will both be running on the same platform and therefore comparable to the PHBs and regular users.
As I understand it, what you are suggesting will work well on an Actor-based system. If I am misunderstanding your intention, please correct me. Most OO languages do not natively support this paradigm, instead allowing several threads of control to invoke methods, making their implementations inherently stack-based. Also, programs written in actor-based languages have been found to be generally much more difficult to maintain and debug than their "regular" counterparts.
Very well, some technical resons then.
on
Free Solaris 8
·
· Score: 1
I admit that it _does_ sound lame to stop using a system due to these reasons. However, these reasons were only the small shove I needed. I use FBSD (previously Linux) at home mainly in a hobby capacity; it is an excellent way to study Unix at home. Coming from a mainly commercial-Unix background background, I was becoming increasingly displeased with the inconsistent feel of Linux. Certainly, RedHat have made serious efforts in giving their distribution a commercial feel, but the cohesion isn't there. That is something that I _did_ find in FreeBSD, despite it being a very different beast to SVR[34] (to which I am accustomed). Now, before you all flame me, I admit that I am not a kernel hacker: I have studied STREAMS and DDI driver development, but besides a couple of basic modules haven't done anything of note in that direction, particularly on a free OS. I do not have the tools to personally evaluate the merits of the Linux or FreeBSD kernels. However, I find that the packaging and organisation of FreeBSD is superior by far to that of the Linux distributions I have used (and closer in quality to that of commercial Unixes), and made me feel "at home" (after a certain acclimatisation period, of course). In the BSD discussion groups I found a helpful and above all _professional_ atmosphere, which I found lacking in those of Linux. Particularly in Slashdot, where the kiddies are out to preach their 3l337ness (sorry Rob & Hemos, you've done an excellent job with your forum; it's a daily spot for me). Posts like this guy's simply infuriate me. Compounded with a general attitude of "well have your SysV's and BSD's, we'll crush you all anyway" from intelligent and professional people I've worked and studied with, I am left disgusted. I'm sorry if I made it seem that this was the reason I've decided to try out an alternative; it was just a trigger. After sampling the news groups and reading about the system, I simply decided to give it a go, and ordered the FreeBSD 3.2 CD's. I am extremely glad that I've switched over. And until the Hurd becomes stable for a serious try-out, I think I'll stick with the BSD's.
Amen brother:) You echo my exact sentiment. The shrill disinformation and plain idiocy that the penguin camp is increasingly becoming full of, is truly disgusting. Being a long-time Unix sysadmin and developer I am very saddened by this. Linux zealots have always been the noisiest ones around, but what we are witnessing is indeed a phenomenon on a Microsoftish scale. I've dumped Linux for FreeBSD about 6 months ago, and haven't looked back. It is very evident that BSD and SVR4 represent more than 2 decades of research, something that many in the linux camp choose to overlook. I believe this will damage the evolution of Linux in the long term. It seems better to them to trash "the competition" than to learn from it.
Re:Ah, time to FUD again.
on
Free Solaris 8
·
· Score: 1
Well well, this is indeed a nicely crafted piece of FUD from the Linux Media 3l337.
> And the brilliant Sun engineers who designed and built those machines find out that their OS counterparts let them down.
Poor, poor hardware engineers. I'm weepy.
> Unless you're one of a handful who need an E10k with dozens of processors, Linux is the only way to go on SPARCs. Try it; you'll never go back.
Do you have benchmarks to back your claims? I'd like to see Linux smoke Solaris on a 2-CPU SPARC box (4 is already way, way too many). Benchmarks? no. Why? Because Linux can't do it. Otherwise the zealots (and ZDnet, you can be sure) would be plastering them all over the net. The Mind$haft benchmark showed one thing, and that is that Linux is good on 1 CPU, useless on more than 2.
> If you amalgamated the very best of all these into one product and called it Linux (for lack of a better name), I doubt you'd find any Solaris in it at all. But that's what you'll get, 10 years from now. And Solaris? A distant memory.
*ROTFL*. Excellent, excellent:] Can I have some of what you're having? Why on earth should a company which has invested hundreds of millions of $ into developing a system tailored to the hardware it's making throw away that product (which is clearly superior and years ahead of the penguin thing) just to follow the cult du jour? Nobody has yet answered this. Why should it "amalgamate" all its diverse products, throw in a bit of penguin dust, and call it Linux, for God's sake? Do you seriously thing IBM is going to dump AIX/Monterey or whatever for Linux?
You guys need to learn to stop and think/check the facts once in a while, but then I guess it's difficult to put down all that youthful exuberance, eh?
What...is...your...problem?
on
Free Solaris 8
·
· Score: 1
...?...?...?
Down, rabid linux puppy, down!!
on
Free Solaris 8
·
· Score: 2
Dear oh dearie me. Yet another zealot talking out of his ass. A couple of remarks:
> Linux runs on a proper superset of the platforms Solaris runs on.
Well done, I see we have done some Set theory at school, haven't we? What is the significance of this? Sun is making most of its money from selling hardware. Sun makes an OS that runs _well_ on this hardware. How this implies "suckiness" is beyond me.
> Solaris is pretty strictly system V, while Linux is some SysV, some BSD, and some "other."
So this mishmash makes Linux more "unsucky". Go figure. FYI, Solaris is SysV _and_ BSD. That's a large part of what Sun brought to the SVR4 table.
> Solaris uses the UFS filesystem. Linux uses primarily the Ext2 filesystem .
This fact actually works _against_Linux; ext2fs (which I guess is probably the only FS you've worked with besides FAT and maybe NTFS) is not half as reliable as UFS. I don't know the performance numbers, but that's irrelevant - reliability is what you're after in an _enterprise_ OS. Solaris also utilises VXFS (if you pay something extra or order a large configuration), which is a full-blown journalling FS allowing smart volume management etc.
> The obvious differences, like licensing...
OK, I see that licensing is what matters to you. Oh well, I won't argue with that, but it seems that it is _mostly_ what matters to you. Well, in that case you are welcome to not use a commercial, non-opensource OS to your heart's content. However, it also seems to me that you are not exactly knowledgeable on any OS other than a certain free one, and therefore are looking to get as many karma points here on slashdot by trolling. You should be moderated below the ground. Let us continue.
> On common hardware, Linux tends to be faster, especially for interactive tasks. > Solaris might be faster on machines with 16 processors or more.
If you are talking about single-CPU workstations, you might be right. But not by much. On any system with 2 or more CPUs, almost _any_ SVR4MP system (e.g. Solaris) smokes Linux. It is a kernel that has been built from the ground up to run on SMP hardware (including the TCP/IP stack), unlike Linux which is undergoing painful and destabilizing surgery to make it be capable of going anywhere near a 4-CPU machine. On Solaris in particular, you can dedicate groups of CPU's to specific processes. And binding a network adapter to a CPU (which Linux is absolutely incapable of) is also possible on most SVR4 systems (haven't done that on Solaris, but I know it is possible on other SVR4's, e.g. UnixWare). Linux has a very long way to go before it can be considered an enterprise server OS. Give it a couple of years - till then, the desktop and a bit file/web serving, no more. It simply doesn't scale.
Hope this puts things slightly more in perspective for you. I wish all these Linux zealots would at least consider getting the facts right next time before entering the discussion. Maybe even rely on some _solid_ experience, for a change? Can you manage this, at least?
*vent* It's exactly this sort of rabidness that's putting off more and more people off this system; I've stopped using Linux and started on FreeBSD some 6 months ago exactly because of this sort of behaviour in the Linux camp, and frankly I'm glad I did it. I've had been a Linux user since 1994 (Slackware then RedHat), and this kind of trolling had been going on since then, only in smaller amounts. Sure there are many responsible users/developers out there, and Linux is definitely a good system, but the signal to noise ratio in all non strictly technical discussions has been growing smaller and smaller by the month. This has indeed become the Cult of the Penguin. */vent*
Mr. Dillon was not throwing mud in order to get people to love him - this is someone who has already earned his status and respect. He was expressing _his opinion_, and being a kernel engineer he has the right to express it. Just as you believe that spawning threads is the right way, the specific branch of OS R&D research they are doing is going another way. The technology that Mr. Dillon is developing has proven itself in many ways, while the NT way of doing things has not, in many ways, and the way their kernel is engineered is, to his opinion, very wrong. The fact that he does not choose to elaborate on his opinion is his choice to make, and it should be respected (again, on the strength of his standing as a pillar of the BSD development community). As he wrote at the end of the piece, it was really much longer than he intended it to be. Now, regarding your quibble: Many kernel architects are convinced that a MK-based design is not fit for a server architecture (which is what MS is trying to pass NT as). Particularly a poorly-implemented one such as the NT microkernel and its bloated set of services. Mr. Dillon may have other complaints about the basic NT architecture, but had he stated only that, it would not be enough for you - you'd demand examples, bring up BeOS etc. It is a fact that MS has had to modify NT's structure to make it up to scratch in the performance department, however even these changes had been poorly architected (video subsystem in the kernel, anyone?). Hopefully Win2K will be a worthy server platform (it seems to be a decent desktop system, at least). Unless MS have gotten their act together, the clean (if extremely slow) basic design of NT will have been further turned into an insecure hodgepodge of services. So, the next time you start flaming, take a deep breath and try to see just what you're doing. Unless, that is, it was your intention to divert attention from the discussion, in which case I am sorry I wasted my breath on you.
Is it because SCSI drives reserve a large portion of their space for alternate tracks (in case there's damage to a track)? In any case, given the number of platters in large HD's, it is easy for the HD makers to remove a couple - with SCSI, you're already expecting to pay more for less space, so they might be milking you just that extra bit. Anyway, if you're a desktop user, there's usually no reason to go SCSI. If you have a server which is used for any non-trivial task (e.g. a bit of print spooling or lightly loaded file server), get SCSI.
You can use XFree86 for your X. Note that if Solaris x86 supports hires VESA modes you might do fine (no way but to try it). Take care with laptops which are more "exotic" with their BIOS structure, e.g. with their peripheral and APM handling.
One other option you could try is to run UnixWare 7.1.1. It generally has decent laptop support, asnd is a full SVR4 (5) system, with CDE,Motif and all the other stuff. Faster than Solaris on x86, and has far better hardware support. You can get an "educational" license from SCO's site. If you can't get a media set you might have to order one from SCO. One gotcha with UnixWare is the while 7.1.1 is coming out now, if you install 7.1.0 you'll have to install something like 30 patches to make it 100% mission critical. If you just need it as a client system (e.g. laptop or desktop), you'll just have to install about 10.
Indeed, the BeOS is a shining example of what a well designed and implemented MK OS is capable of; excellent response, easy (and fun) to write for. Having seen the QNX demo diskette run on an old 486DX, I'd say that system is more proof that MK's can be used as the basis for a fast and stable system. However, the great performance problem with MK OS's becomes most apparent when they are heavily loaded; message-passing (and to a lesser extent context switching) takes a bigger and bigger slice of the total system's running time, which means you get effectively much less work. That's where monolithic systems (e.g. Unix) greatly outpace the MK ones. I'd say that MK systems are probably more well-suited to workstations and monolithic systems to servers.
Well, if you're into Unix-on-Intel benchmarks, UnixWare 7 is still fastest and most scalable overall according to any benchmark. I'm quite sure that a carefully tuned and properly recompiled Linux kernel will do as well in some respects (if not better) on a 1-CPU system, however add a couple of CPU's (or boxen - check out NonStop Clustering) and you've blown the competition away. Real life isn't always about what benchmark you've used. However, this system is tailored to Intel hardware, while Solaris isn't (and never will be). The Solaris way of doing things may be more to your taste, however Sun will not optimize its performance on Intel hardware, for obvious reasons. I'm sure that it is a very stable system, but it will never be a particularly swift one.
And while we're at it, check these out:
http://www.sco.com/press/releases/1999/6886.html
http://www.sco.com/press/releases/1998/6836.html
http://www.sco.com/press/releases/1998/6822.html
Particularly the last one. I don't really find the second one (SPEC_int/fp to measure cluster scalability) very applicable to a large server environment, but what the hell - if anyone prefers NSC over Beowulf/Mosix to do physics research, they probably deserve it:)
Isn't it great the OSS movement is giving these guys PR ammo? Oh well, at least it seems Linux is benefitting as well; it's all about mindshare. In any case, you can be sure there won't be much Linux running on these boxen, even in VM's. AIX/Monterey is what will be running there - it's optimised to do that, and will (does?) run PPC binaries natively (see lxrun).
Sure, as a desktop FreeBSD user this would make me happy, as it would boost FreeBSD acceptance. However, I'd rather see HURD ready first (this is one of Debian's projects right now). It seems like a very interesting and worthwhile alternative to current kernels, and might show the world that the OSS movement can produce a true, quality microkernel-based system.
The message SGI is giving the market is indeed that IRIX may be sufficient now, but with a bit of effort Linux wiull probably surpass it. Anyone who'd been considering buying SGI will now go to someone who's 100% behind their systems (and I'm sure there will be many current customers who'll shift because of this as well). Idiots.
Note that what you are proposing is not all that simple: _increasing modularity_ and reworking the _SMP_ scalability of the kernel invovles _major changes_ to the kernel structure. Here lies a great risk - that the changes will be so sweeping, that the linux kernel structure will fragment: we'll have the "Linus" kernel, and the "SGI" kernel. Hopefully if this happens, the changes will merge into Linus's kernel quickly, otherwise if SGI release their changes to quickly we'll have irreversible fragmentation.
Slashdot Mirror
Alpha has had this probably from the beginning (at least since '94, when I saw this in an Alpha ISA document). I also recall seeing prefetch instructions on at least one VLIW design. Do (some) L2 caches enable the CPU to issue line/block prefetch commands?
What about cross-platform Smalltalk tools (e.g. ST/X, IBM?)?
That's pretty weak...
>> If you don't need MS Windows to run MS office, then you don't need MS Visual Studio to write apps with...
Now _here's_ a terrifying thought - what if porting Office necessitates porting a part of MFC => Unix cross development on Visual Studio. Then you get Office for Linux, IE for Linux etc. _and_ a development environment from MS (even if they only use that in-house). This means that MS put their browser into Linux/Unix-space; now here's some serious competition for any browser on Unix (including Mozilla for sure). I wonder what kind of havoc that will wreak...
I for one am looking forward to having Office on Unix - even if the product sucks as much as it does now, it will be common ground for replacement of Windows boxes with Unix. Then we have the possibility of migrating from MS Office to Star Office or whatever, since they will both be running on the same platform and therefore comparable to the PHBs and regular users.
As I understand it, what you are suggesting will work well on an Actor-based system. If I am misunderstanding your intention, please correct me.
Most OO languages do not natively support this paradigm, instead allowing several threads of control to invoke methods, making their implementations inherently stack-based. Also, programs written in actor-based languages have been found to be generally much more difficult to maintain and debug than their "regular" counterparts.
LOL.
I admit that it _does_ sound lame to stop using a system due to these reasons. However, these reasons were only the small shove I needed.
I use FBSD (previously Linux) at home mainly in a hobby capacity; it is an excellent way to study Unix at home.
Coming from a mainly commercial-Unix background background, I was becoming increasingly displeased with the inconsistent feel of Linux. Certainly, RedHat have made serious efforts in giving their distribution a commercial feel, but the cohesion isn't there. That is something that I _did_ find in FreeBSD, despite it being a very different beast to SVR[34] (to which I am accustomed).
Now, before you all flame me, I admit that I am not a kernel hacker: I have studied STREAMS and DDI driver development, but besides a couple of basic modules haven't done anything of note in that direction, particularly on a free OS. I do not have the tools to personally evaluate the merits of the Linux or FreeBSD kernels. However, I find that the packaging and organisation of FreeBSD is superior by far to that of the Linux distributions I have used (and closer in quality to that of commercial Unixes), and made me feel "at home" (after a certain acclimatisation period, of course).
In the BSD discussion groups I found a helpful and above all _professional_ atmosphere, which I found lacking in those of Linux. Particularly in Slashdot, where the kiddies are out to preach their 3l337ness (sorry Rob & Hemos, you've done an excellent job with your forum; it's a daily spot for me). Posts like this guy's simply infuriate me. Compounded with a general attitude of "well have your SysV's and BSD's, we'll crush you all anyway" from intelligent and professional people I've worked and studied with, I am left disgusted. I'm sorry if I made it seem that this was the reason I've decided to try out an alternative; it was just a trigger. After sampling the news groups and reading about the system, I simply decided to give it a go, and ordered the FreeBSD 3.2 CD's. I am extremely glad that I've switched over. And until the Hurd becomes stable for a serious try-out, I think I'll stick with the BSD's.
Amen brother :)
You echo my exact sentiment. The shrill disinformation and plain idiocy that the penguin camp is increasingly becoming full of, is truly disgusting. Being a long-time Unix sysadmin and developer I am very saddened by this. Linux zealots have always been the noisiest ones around, but what we are witnessing is indeed a phenomenon on a Microsoftish scale.
I've dumped Linux for FreeBSD about 6 months ago, and haven't looked back. It is very evident that BSD and SVR4 represent more than 2 decades of research, something that many in the linux camp choose to overlook. I believe this will damage the evolution of Linux in the long term. It seems better to them to trash "the competition" than to learn from it.
Well well, this is indeed a nicely crafted piece of FUD from the Linux Media 3l337.
:] Can I have some of what you're having?
> And the brilliant Sun engineers who designed and built those machines find out that their OS counterparts let them down.
Poor, poor hardware engineers. I'm weepy.
> Unless you're one of a handful who need an E10k with dozens of processors, Linux is the only way to go on SPARCs. Try it; you'll never go back.
Do you have benchmarks to back your claims? I'd like to see Linux smoke Solaris on a 2-CPU SPARC box (4 is already way, way too many). Benchmarks? no. Why? Because Linux can't do it. Otherwise the zealots (and ZDnet, you can be sure) would be plastering them all over the net. The Mind$haft benchmark showed one thing, and that is that Linux is good on 1 CPU, useless on more than 2.
> If you amalgamated the very best of all these into one product and called it Linux (for lack of a better name), I doubt you'd find any Solaris in it at all. But that's what you'll get, 10 years from now. And Solaris? A distant memory.
*ROTFL*. Excellent, excellent
Why on earth should a company which has invested hundreds of millions of $ into developing a system tailored to the hardware it's making throw away that product (which is clearly superior and years ahead of the penguin thing) just to follow the cult du jour? Nobody has yet answered this. Why should it "amalgamate" all its diverse products, throw in a bit of penguin dust, and call it Linux, for God's sake? Do you seriously thing IBM is going to dump AIX/Monterey or whatever for Linux?
You guys need to learn to stop and think/check the facts once in a while, but then I guess it's difficult to put down all that youthful exuberance, eh?
...?...?...?
Dear oh dearie me. Yet another zealot talking out of his ass.
...
A couple of remarks:
> Linux runs on a proper superset of the platforms Solaris runs on.
Well done, I see we have done some Set theory at school, haven't we?
What is the significance of this? Sun is making most of its money from selling hardware. Sun makes an OS that runs _well_ on this hardware. How this implies "suckiness" is beyond me.
> Solaris is pretty strictly system V, while Linux is some SysV, some BSD, and some "other."
So this mishmash makes Linux more "unsucky". Go figure.
FYI, Solaris is SysV _and_ BSD. That's a large part of what Sun brought to the SVR4 table.
> Solaris uses the UFS filesystem. Linux uses primarily the Ext2 filesystem .
This fact actually works _against_Linux; ext2fs (which I guess is probably the only FS you've worked with besides FAT and maybe NTFS) is not half as reliable as UFS. I don't know the performance numbers, but that's irrelevant - reliability is what you're after in an _enterprise_ OS. Solaris also utilises VXFS (if you pay something extra or order a large configuration), which is a full-blown journalling FS allowing smart volume management etc.
> The obvious differences, like licensing
OK, I see that licensing is what matters to you. Oh well, I won't argue with that, but it seems that it is _mostly_ what matters to you. Well, in that case you are welcome to not use a commercial, non-opensource OS to your heart's content. However, it also seems to me that you are not exactly knowledgeable on any OS other than a certain free one, and therefore are looking to get as many karma points here on slashdot by trolling. You should be moderated below the ground.
Let us continue.
> On common hardware, Linux tends to be faster, especially for interactive tasks.
> Solaris might be faster on machines with 16 processors or more.
If you are talking about single-CPU workstations, you might be right. But not by much.
On any system with 2 or more CPUs, almost _any_ SVR4MP system (e.g. Solaris) smokes Linux. It is a kernel that has been built from the ground up to run on SMP hardware (including the TCP/IP stack), unlike Linux which is undergoing painful and destabilizing surgery to make it be capable of going anywhere near a 4-CPU machine. On Solaris in particular, you can dedicate groups of CPU's to specific processes. And binding a network adapter to a CPU (which Linux is absolutely incapable of) is also possible on most SVR4 systems (haven't done that on Solaris, but I know it is possible on other SVR4's, e.g. UnixWare).
Linux has a very long way to go before it can be considered an enterprise server OS. Give it a couple of years - till then, the desktop and a bit file/web serving, no more. It simply doesn't scale.
Hope this puts things slightly more in perspective for you. I wish all these Linux zealots would at least consider getting the facts right next time before entering the discussion. Maybe even rely on some _solid_ experience, for a change? Can you manage this, at least?
*vent*
It's exactly this sort of rabidness that's putting off more and more people off this system; I've stopped using Linux and started on FreeBSD some 6 months ago exactly because of this sort of behaviour in the Linux camp, and frankly I'm glad I did it. I've had been a Linux user since 1994 (Slackware then RedHat), and this kind of trolling had been going on since then, only in smaller amounts. Sure there are many responsible users/developers out there, and Linux is definitely a good system, but the signal to noise ratio in all non strictly technical discussions has been growing smaller and smaller by the month. This has indeed become the Cult of the Penguin.
*/vent*
Mr. Dillon was not throwing mud in order to get people to love him - this is someone who has already earned his status and respect. He was expressing _his opinion_, and being a kernel engineer he has the right to express it. Just as you believe that spawning threads is the right way, the specific branch of OS R&D research they are doing is going another way. The technology that Mr. Dillon is developing has proven itself in many ways, while the NT way of doing things has not, in many ways, and the way their kernel is engineered is, to his opinion, very wrong.
The fact that he does not choose to elaborate on his opinion is his choice to make, and it should be respected (again, on the strength of his standing as a pillar of the BSD development community). As he wrote at the end of the piece, it was really much longer than he intended it to be.
Now, regarding your quibble: Many kernel architects are convinced that a MK-based design is not fit for a server architecture (which is what MS is trying to pass NT as). Particularly a poorly-implemented one such as the NT microkernel and its bloated set of services. Mr. Dillon may have other complaints about the basic NT architecture, but had he stated only that, it would not be enough for you - you'd demand examples, bring up BeOS etc.
It is a fact that MS has had to modify NT's structure to make it up to scratch in the performance department, however even these changes had been poorly architected (video subsystem in the kernel, anyone?). Hopefully Win2K will be a worthy server platform (it seems to be a decent desktop system, at least). Unless MS have gotten their act together, the clean (if extremely slow) basic design of NT will have been further turned into an insecure hodgepodge of services.
So, the next time you start flaming, take a deep breath and try to see just what you're doing. Unless, that is, it was your intention to divert attention from the discussion, in which case I am sorry I wasted my breath on you.
Is it because SCSI drives reserve a large portion of their space for alternate tracks (in case there's damage to a track)?
In any case, given the number of platters in large HD's, it is easy for the HD makers to remove a couple - with SCSI, you're already expecting to pay more for less space, so they might be milking you just that extra bit.
Anyway, if you're a desktop user, there's usually no reason to go SCSI. If you have a server which is used for any non-trivial task (e.g. a bit of print spooling or lightly loaded file server), get SCSI.
Get on it.
That's what the source is there for.
You can use XFree86 for your X. Note that if Solaris x86 supports hires VESA modes you might do fine (no way but to try it).
Take care with laptops which are more "exotic" with their BIOS structure, e.g. with their peripheral and APM handling.
One other option you could try is to run UnixWare 7.1.1. It generally has decent laptop support, asnd is a full SVR4 (5) system, with CDE,Motif and all the other stuff. Faster than Solaris on x86, and has far better hardware support.
You can get an "educational" license from SCO's site. If you can't get a media set you might have to order one from SCO.
One gotcha with UnixWare is the while 7.1.1 is coming out now, if you install 7.1.0 you'll have to install something like 30 patches to make it 100% mission critical. If you just need it as a client system (e.g. laptop or desktop), you'll just have to install about 10.
I'm certain he never meant to offend your sensibilities in such a severe manner. Please be a tad more respectful.
This is a wonderful piece of work. I don't think I've seen a post with so many contradictory statements on Slashdot yet :)
Indeed, the BeOS is a shining example of what a well designed and implemented MK OS is capable of; excellent response, easy (and fun) to write for. Having seen the QNX demo diskette run on an old 486DX, I'd say that system is more proof that MK's can be used as the basis for a fast and stable system.
However, the great performance problem with MK OS's becomes most apparent when they are heavily loaded; message-passing (and to a lesser extent context switching) takes a bigger and bigger slice of the total system's running time, which means you get effectively much less work. That's where monolithic systems (e.g. Unix) greatly outpace the MK ones.
I'd say that MK systems are probably more well-suited to workstations and monolithic systems to servers.
Well, if you're into Unix-on-Intel benchmarks, UnixWare 7 is still fastest and most scalable overall according to any benchmark. I'm quite sure that a carefully tuned and properly recompiled Linux kernel will do as well in some respects (if not better) on a 1-CPU system, however add a couple of CPU's (or boxen - check out NonStop Clustering) and you've blown the competition away.
l
l
l
:)
Real life isn't always about what benchmark you've used. However, this system is tailored to Intel hardware, while Solaris isn't (and never will be). The Solaris way of doing things may be more to your taste, however Sun will not optimize its performance on Intel hardware, for obvious reasons. I'm sure that it is a very stable system, but it will never be a particularly swift one.
And while we're at it, check these out:
http://www.sco.com/press/releases/1999/6886.htm
http://www.sco.com/press/releases/1998/6836.htm
http://www.sco.com/press/releases/1998/6822.htm
Particularly the last one.
I don't really find the second one (SPEC_int/fp to measure cluster scalability) very applicable to a large server environment, but what the hell - if anyone prefers NSC over Beowulf/Mosix to do physics research, they probably deserve it
Isn't it great the OSS movement is giving these guys PR ammo? Oh well, at least it seems Linux is benefitting as well; it's all about mindshare.
In any case, you can be sure there won't be much Linux running on these boxen, even in VM's. AIX/Monterey is what will be running there - it's optimised to do that, and will (does?) run PPC binaries natively (see lxrun).
Excellently said, my hat off to you, my friend!!!
Sure, as a desktop FreeBSD user this would make me happy, as it would boost FreeBSD acceptance. However, I'd rather see HURD ready first (this is one of Debian's projects right now). It seems like a very interesting and worthwhile alternative to current kernels, and might show the world that the OSS movement can produce a true, quality microkernel-based system.
Whatever...
The message SGI is giving the market is indeed that IRIX may be sufficient now, but with a bit of effort Linux wiull probably surpass it. Anyone who'd been considering buying SGI will now go to someone who's 100% behind their systems (and I'm sure there will be many current customers who'll shift because of this as well). Idiots.
Note that what you are proposing is not all that simple: _increasing modularity_ and reworking the _SMP_ scalability of the kernel invovles _major changes_ to the kernel structure. Here lies a great risk - that the changes will be so sweeping, that the linux kernel structure will fragment: we'll have the "Linus" kernel, and the "SGI" kernel. Hopefully if this happens, the changes will merge into Linus's kernel quickly, otherwise if SGI release their changes to quickly we'll have irreversible fragmentation.
"...no need to invoke time travelling people from the future when perfectly good genetic defects will do"
Can I quote this?