That's the trouble with using a server chip, you get useless features like multiple OS support, instead of focussing on things like ramping up the megahertz.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Next PowerPC To Support Multiple OS's?
- Thread starter MacRumors
- Start date
- Sort by reaction score
If this was implemented on systems that support it, here's how it would work:JonMaker said:UPDATES
In the server world (and even at home for me) rebooting to install low level OS updates is a big deal. IT staff can be fired for five minutes of downtime.
Would this multi OS thingy allow stuff like hot-swapping the kernel?
e.g. Would we be able to boot one OS, update it (after a few months of uptime), then "hot-reboot" into the newly updated OS?
1. You run the old system, and download and install the updater.
2. After the updater installs, the updated OS boots silently in another partition.
3. You restart like you normally would, except what ends up happening is you get switched over to the updated OS while the old OS is shut down.
4. Once the old OS shuts down completely, its components that are still around that were updated get deleted, and you can continue working normally.
AidenShaw said:
First of all I did not claim NT was stuck to x86. A claimed every version AFTER NT got more and more stuck to x86, because they dropped the multi platform thing. Read again before telling it's an ignorant statement.
It's a FACT that they dropped MIPS, PPC and Alpha. And why on earth would they waste resources in still maintaining MIPS, PPC and Alpha (note: 2 DEAD processors in that list) ports of their products???
And yes I know about HAL, but if I'm not mistaken, there is no such thing in XP, or at least they would have changed its name because it wouldn't be appropriate anymore. And yes I also know W2K/Alpha, in fact I own a copy.
Does that even matter anymore? The HAL is still around in Windows XP, and it's still called HAL. The purpose it serves now is probably less cross-platform and more with things like allowing the OS to use the graphics card for certain tasks regardless of what kind of graphics card it is.MacNeXT said:
Mainly Useful for Servers
This would mainly be useful for virtual servers, and may not have any relevance to the chips Apple would buy.
Macrumors said:
This would mainly be useful for virtual servers, and may not have any relevance to the chips Apple would buy.
Future of computing?
This might come in handy for real power users or people wanting to operate a two-in-one Mac simultaniously with tasks exceeding one core's power but all in all I don't find it useful for avarage users, at least not at this moment.
This might come in handy for real power users or people wanting to operate a two-in-one Mac simultaniously with tasks exceeding one core's power but all in all I don't find it useful for avarage users, at least not at this moment.
My Dell laptop already has a dual-core MIPS chip....
The "tigon" series of gigabit ethernet chips use a pair of MIPS R4000 embedded cores to do TCP offloading from the host CPU, and other tasks.
The "tigon3" is manufactured by Broadcom and used in a number of different cards. The chip is the BCM570x series, and is frequently found as the embedded GbE controller on systems and laptops from major and minor companies.
Eric_Z said:
The "tigon" series of gigabit ethernet chips use a pair of MIPS R4000 embedded cores to do TCP offloading from the host CPU, and other tasks.
The "tigon3" is manufactured by Broadcom and used in a number of different cards. The chip is the BCM570x series, and is frequently found as the embedded GbE controller on systems and laptops from major and minor companies.
Tuttle said:
BS. Their ThinkPad line is a top seller among laptop brands. Many of the Fortune 500 companies of the world use..or at least did use..IBM. Desktops are another matter. I know of no one who uses IBM desktops.
This delusional thinking that the PC world is going to dry up and go belly up is more Apple cool aid being handed around. PC sales are strong however they are being cannibalized by companies like Dell that sell their wares for bottom of the barrel, for fair to inferior quality, that quality players like IBM can't keep up. Dell makes their money in quantity, corp sales, accessories, and extended warrantees. I suppose if IBM wanted to make cheap crap they could compete against Dell. But why?
~Shard~ said:I don't know if I'd want Windows anywhere near my OS X though.
I don't know... I think I would love to have the option of running windows on a mac.
It would meet all my needs at school (that I currently have to use a PC for),
It would thrill my son who loves to play on-line games (many of which he can't play well on a pc) and allow him to download those obscure apps that he and his friends love to play with (that are only available on PC)
From my perspective this removes issues that people may have with buying a mac. If you can have the ease of use and elegance of using a mac without missing any of the pc only apps people it may create a demand for the pc only apps to be ported.
Can't be closed minded about this (as much as I would like to be) the world uses Windows.
MacNeXT said:
I don't need to read it again, because you have again repeated the mistake.
PPC and MIPS dropped out early - no sales. The Alpha port stayed through the Windows 2000 beta and release candidate stages, in both 32-bit (as in Windows NT 4 Alpha) and the new 64-bit Win64 version that was in development.
Windows 2000 was released on x86 and IA64, two very different architectures. Today XP is running on x86, IA64 and x64 - same with Server 2003. How is that "more stuck" on x86? How is that "dropping multi-platform"???
MacNeXT said:
The Win64 version continued (after Alpha was dropped from Win2K) to be built and used internally for 64-bit development - MS had a lot of Alpha systems for 64-bit work, and very few IA64 systems at the time.
The reason to "waste resources" is quite simple - compiling the other architectures is a sanity check to ensure that the code does not get "stuck on x86". It's mostly a compute cost, not a major people effort.
Plus, it's insurance. You never know when you might want to run NT on a PowerMac G5 (http://www.gamesindustry.biz/content_page.php?section_name=dev&aid=3039)
MacNeXT said:
HAL is the only way to touch the hardware - even device drivers have to go through the HAL. There are different HALs for single CPU and multi-CPU systems, and different power management architectures.
Your Win2K/Alpha is a beta 32-bit version, right? Alpha was dropped as a target before the official release of Win2K, and the 64-bit versions were mainly used by developers inside MS and Digital/Compaq.
.done deal.
There's already an IA64 HAL series and HALs for the x64 (EM64T and AMD64).
And they're not "retrofits" - the IA64 HAL was part of Windows 2000 for Itanium, and the x64 one is just a new piece to support the low-level architecture of the Athlon 64/Opteron/Xeon-64 systems.
The HAL is actually a virtualization layer - it deals with hardware differences so that the rest of NT can run on an idealized virtual platform.
Device drivers don't have to deal with different bus layouts - the HAL hides the details behind a common API.
The HAL does not change the instructions though - all of the other NT code has to be recompiled for IA64 (or PPC or MIPS or Alpha).
x64 is a little different - some things can stay 32-bit (the "notepad" editor, for example) under x64. All 64-bit code and supporting libraries would need to be recompiled.
You'd also want to recompile any performance sensitive code, since x64 64-bit is usually faster than x86 32-bit.
MacNeXT said:
There's already an IA64 HAL series and HALs for the x64 (EM64T and AMD64).
And they're not "retrofits" - the IA64 HAL was part of Windows 2000 for Itanium, and the x64 one is just a new piece to support the low-level architecture of the Athlon 64/Opteron/Xeon-64 systems.
The HAL is actually a virtualization layer - it deals with hardware differences so that the rest of NT can run on an idealized virtual platform.
Device drivers don't have to deal with different bus layouts - the HAL hides the details behind a common API.
The HAL does not change the instructions though - all of the other NT code has to be recompiled for IA64 (or PPC or MIPS or Alpha).
x64 is a little different - some things can stay 32-bit (the "notepad" editor, for example) under x64. All 64-bit code and supporting libraries would need to be recompiled.
You'd also want to recompile any performance sensitive code, since x64 64-bit is usually faster than x86 32-bit.
Multicore OSX?
I personally think that IF Apple ever uses multiOS feature of those chips, they'll use it for their own multicore OS, not for providing multiple OS-s on their machines. May happen a OSX with one DArwin and other Linux core (for handling Linux binaries natively).
But it's far more likely, that such feature won't be used.
I personally think that IF Apple ever uses multiOS feature of those chips, they'll use it for their own multicore OS, not for providing multiple OS-s on their machines. May happen a OSX with one DArwin and other Linux core (for handling Linux binaries natively).
But it's far more likely, that such feature won't be used.
This is nothing new for IBM
There's really nothing that new here. IBM mainframes can already run multiple instances of Linux on the same machine. As another poster already mentioned, the idea is to partition a bigger machine into smaller "virtual" machines. The benefit of having multiple "virtual" machines instead of multiple physical machines would be purely cost and easier maintenance, I think. A desktop user would have almost no use for this.
There's really nothing that new here. IBM mainframes can already run multiple instances of Linux on the same machine. As another poster already mentioned, the idea is to partition a bigger machine into smaller "virtual" machines. The benefit of having multiple "virtual" machines instead of multiple physical machines would be purely cost and easier maintenance, I think. A desktop user would have almost no use for this.
I think that this is very good news. IBM is working hard so that Apple will be able to keep up with the rest of technology. It would seem to be much better to be able to run Windows natively on the Mac. No Virtual PC to slow down the process. If it is done by Apple it will always be better.
MacNeXT said:
Don't forget, Apple released a severely hacked version of the FreeBSD core that is nowhere near as stable as UNIX: there are plenty of power users that lock up their systems all the time with virtual memory traps and hardware freezes that should never (and i do mean NEVER) freeze a unix box. All those deck problems you see running Final Cut? Every piece of terrible arbitrary software that Avid/Digidesigns releases? Video driver issues? Printer crashes? Network errors? All of those come back down to problems at the core level of the operating system which, if it were truly properly written, OS X could run unaffected by.
As for the Linux users not seeing the mac os as an appealing platform, keep in mind that while it's a hacked and buggy version of unix, the mac os still serves as a version of unix that has graphics capabilities. this alone is important enough since everyone wants to be able to use quicktime and windows media (and sometimes little programs like photoshop) no matter what platform they run on. Being able to run two computers in one box has serious benefits if it's done with hardware instead of with software because it removes programming errors from the chart along with providing much faster system performance. trust me, the only way this could be better news for apple is if someone starts placing 4 of these in machines instead of only 2...
What reason do I have to believe that the core of Mac OS X is "hacked"? Right now, none at all. I agree that it's buggy for the simple reason that it's next to impossible for even tiny software programs to be bug-free.alfismoney said:
As we all know, Window boxes outnumber Mac boxes big-time.
Many of us Mac users have bought Virtual PC so that we can run Windows applications when the need arises.
Those of us who run Virtual PC are aware of VPC's massive performance hit on ram and processor speed.
It would be nice to run a Windows application without the performance hit of Virtual PC.
I look forward to a new processor chip that could run Windows and Mac OS at the same time. Eliminate the need for Virtual PC.
Time will tell whether or not the new chip supports Windows and Mac OS, or simply a variety of flavors of the same OS.
Many of us Mac users have bought Virtual PC so that we can run Windows applications when the need arises.
Those of us who run Virtual PC are aware of VPC's massive performance hit on ram and processor speed.
It would be nice to run a Windows application without the performance hit of Virtual PC.
I look forward to a new processor chip that could run Windows and Mac OS at the same time. Eliminate the need for Virtual PC.
Time will tell whether or not the new chip supports Windows and Mac OS, or simply a variety of flavors of the same OS.
how about a realtime version of OSX
A realtime version of OSX would be very useful in science, engineering, and various industrial control and process control systems. A very small market probably, but, it all adds up.
A realtime version of OSX would be very useful in science, engineering, and various industrial control and process control systems. A very small market probably, but, it all adds up.
As a long time Mac user (1989), I almost fear being able to use Windows on my Mac machine. I once had the PC card in an old PowerPC I had, and of course it was terrible (slow, DOS only, etc.). Then years later I bought VPC and last year got VPC 6 - so, like someone else mentioned, I could run some PC apps occasionally. But VPC is such a beast of burden on any Mac, even with 3 gb ram and dual processors. I would certainly love to throw it away, but having Windows run directly without emulation, as this processor may provide, brings me to wonder whether or not viruses written for M$ would start affecting OS X files, or HD structure in general. That's what scares me.
Call me crazy, but didn't Apple once try to create an OS that ran Mac and Windows files simultaneously? Rhapsody I think was the CODENAME for the project, not the name of the OS. Anyone remember that, or am I crazy?
Call me crazy, but didn't Apple once try to create an OS that ran Mac and Windows files simultaneously? Rhapsody I think was the CODENAME for the project, not the name of the OS. Anyone remember that, or am I crazy?
Nice way to avoid porting
One of the problems with porting Unix/Linux apps to OSX is that the OSX implementation of Unix standards is incomplete or broken in places. This makes porting a real pain for any software that isn't vanilla stdlib, and OSX breaks a surprising amount of Unix software in unpleasant ways. It is one of the most "incomplete" common Unix variants out there right now. On the other hand, it is very easy to port other vaguely Unix-y things to Linux because Linux tends to support a superset of every obscure Unix standard or API ever implemented these days.
I would love to be able to run Linux in parallel on the PPC hardware so that I could get a better and more complete Unix environment than OSX provides. For most things, I don't even bother to port to OSX any more even though the code base is server-side Linux C/C++, as it is not worth the effort (especially since the most common target is Linux on AMD64 anyway).
MacNeXT said:
One of the problems with porting Unix/Linux apps to OSX is that the OSX implementation of Unix standards is incomplete or broken in places. This makes porting a real pain for any software that isn't vanilla stdlib, and OSX breaks a surprising amount of Unix software in unpleasant ways. It is one of the most "incomplete" common Unix variants out there right now. On the other hand, it is very easy to port other vaguely Unix-y things to Linux because Linux tends to support a superset of every obscure Unix standard or API ever implemented these days.
I would love to be able to run Linux in parallel on the PPC hardware so that I could get a better and more complete Unix environment than OSX provides. For most things, I don't even bother to port to OSX any more even though the code base is server-side Linux C/C++, as it is not worth the effort (especially since the most common target is Linux on AMD64 anyway).
As far as everyones great speculation that somehow till will enable windows to run on a MAC at a hardware level. I think those hopes are far fetched (but something that I think that more mac users would like than dislike). However the only way I can see this happening is either somehow building a second instruction set on the chip (x86), or more probably (and slower) some sort of on-chip, hardware based, instruction translation to 1 or more PPC instructions. Either way I doubt this what they meant by running multiple OS's. I think you're either looking @ functionality geared towards servers, or Linux/OSX combination with possibilities for R&D envirornments where a variety of computers are used (at work we have every type of desktop you could imagine).
Legacy Sound Card
Wow, this is great news, now I will be able to use my legacy Korg Oasys soundcard with its OS9 drivers whilst using Logic 7 on the same machine!!
Any ideas of a release date??
Dont suppose there's any chance of having this feature in the G5 PBs at some point in the future?
Wow, this is great news, now I will be able to use my legacy Korg Oasys soundcard with its OS9 drivers whilst using Logic 7 on the same machine!!
Any ideas of a release date??
Dont suppose there's any chance of having this feature in the G5 PBs at some point in the future?