The new programming interface introduced in Windows “Longhorn” is called WinFX. Learn the object-oriented, managed API for the breadth of “Longhorn” and how it relates to the work you are already doing on the .NET Framework. Elsewhere, Steve Jobs is speaking of Longhorn and compares it to the Mac OS X long road to maturity and forsees that the same steep road will be there for Microsoft as well.
What is WinFX; Steve Jobs on Longhorn
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
I tried to watch MSDN TV with the new Media Player 9 for Mac (read
the previous article) but Netscape 4.7 is required. After activating the
Debug menu in Safari and forcing the emulation, a missing audio codec
is required…
Microsdt should better evangelize .NET to potential Mac switchers.
What are you talking about? People that leave M$ are on the fringe. I have been using M$ products since 1987 (DOS 3.3). I have seen this company grow and devour and abuse other software companies to no end. I am very tired of M$ and their need to know everything about every user that uses their operating system.
Side note; does anyone remember when manuals actually came with their software? I heard 10.3 comes with 1000+ pages of docs. But alas, I am not a mac user. ( Linux/*BSD user).
I like how the M$ folks dismiss people that use other OS’ and say their on the edge/fringe. Perhaps were all drifters (and what not)?
Please, wake up and smell the Moca/Latte/Mt.Dew.
I thought it sounded like the guy was talking about switchers from Mac to windows.
“Side note; does anyone remember when manuals actually came with their software? I heard 10.3 comes with 1000+ pages of docs. But alas, I am not a mac user. ( Linux/*BSD user). ”
I remember those days. 10.3, like all previous versions of OS X, comes with a picture book of about 20 pages. Simplicity is the key I guess. Now FCP on the other hand comes with a stack of books, and they don’t even cover SoundTrack, LiveType, Compressor, or Cimena Tools, just FCP.
Whats FCP???
FCP is Apples video editing software.
FCP = Final Cut Pro
MS will have just as hard a time, if not harder moving over to longhorn.
I am sure that rather than have the underlying APIs exposed to legacy code that uses them, it will run a managed code legacy interface much like Classic mode in OS X that will expose the APIs in a safe sandbox.
then there is the HUGE problem with Longhorn not running on older hardware and the fact that most windows users run windows 98, so app writers will have to some how figure out how to support the people who are not upgrading because they like the machine they have or they do not want to change OSs since it is working fine for them.
Apple went through the same problem, but at least with apple. the OS could run on old hardware and run pretty well so almost all OS 9 users could upgrade their machines to OS X.
I am sure that rather than have the underlying APIs exposed to legacy code that uses them, it will run a managed code legacy interface much like Classic mode in OS X that will expose the APIs in a safe sandbox.
That is what they do now using WOW (Windows on Windows) which enables one to run crappy win16 and non-pure win32 applications on Windows XP.
Sure, it isn’t an eligant solution but it kind of works š
then there is the HUGE problem with Longhorn not running on older hardware and the fact that most windows users run windows 98, so app writers will have to some how figure out how to support the people who are not upgrading because they like the machine they have or they do not want to change OSs since it is working fine for them.
Microsoft won’t have a problem if they team up with OEM vendors and offer and incentive for users to upgrade, something like, “buy a PC with longhorn pre-installed and get Office Standard for only $20!”.
Microsoft needs to break compatibility BUT they’re unwilling to part with money to “subsidise” the initial longhorn adotion. If they can say to vendors, “here is some money, produce a pure longhorn version of your application” then subsidise a competitive upgrade scheme, “upgrade from win32 version to longhorn for only $20!”.
Microsoft needs to create that base for customers to move foward. Most users aren’t worried about pissant shareware rubbish but if they can’t get the applications they want on the new platform, they’ll stick with “what works” for them.
“Microsdt should better evangelize .NET to potential Mac switchers.”
Why? Apple already has something a lot better and the technology’s been around for a while, it’s called Cocoa.
“the only switchers that you might see are those who are on the very fringe. the rest who would be interested in Longhorn are already running both OSs. I would say 1/2% of Mac users would actually leave the platform and dump Mac totally.”
You make it sound like most Mac OS X users are unhappy with the OS. Why so we can go back to a poorly designed OS? I haven’t been able to find anyone who thinks that Longhorn has a good design going for it. Plus MS thinks you’ll be using 4300×2000 some resolution by 2006 which is totally unrealistic. Every Mac user I’ve seen doesn’t think Longhorn is a huge improvement.
Maybe it’s an improvment from a Windows user perspective, but from anyone using Mac OS X it’s not anything special. Microsoft is essentially catching up to where some Mac OS X features have been for some time now. I don’t know why some people can’t admit it, maybe your jealous or just ignorant? I don’t know or care, Mac OS X Panther 10.3 is already the most advanced & easiest to use unix based os out there and by the time Longhorn comes out mac os x will have advanced even further.
Personally I don’t get why people are all thinking Longhorn & .NET are like ultra special features. I chose to make an upgrade from a good AMD PC w/ Windows XP PRO to a good PowerMac G4 & Mac OS X and I haven’t regretted it since. If you think any Mac users are going to switch because of Longhorn or .NET it’s just wishful thinking.
How many Win98 users will there be in 2006? How many of those users will hold out for whatever is after Longhorn? Will computer techs remember how Win98 works?
then there is the HUGE problem with Longhorn not running on older hardware and the fact that most windows users run windows 98, so app writers will have to some how figure out how to support the people who are not upgrading because they like the machine they have or they do not want to change OSs since it is working fine for them.
Most people are probably like my dad – gonna stick with his P3-600 for a couple more years until Longhorn is released. Until then, there’s really no reason to upgrade. But I doubt most people would stick with Windows 98 beyond 2006 and Longhorn, cuz you know the windows updates will have stopped by then.
BTW: Steve Jobs like to talk a lot of shit, doesn’t he ?
As for OSX, I don’t think it’ll make much of a bigger dent than it already has until/unless it is ported to x86 hardware.
I gave OSX (pre 10.3) a few test drives at Frys and was rather unimpressed, though some people say you have to use it for awhile to really appreciate it (?) … but the question is, how are we supposed to do that if it won’t run on our hardware? I’m not going to go out and blow a couple of grand on one when I wasn’t blown away by the in-store demo.
Even if Apple were to move to an x86 processor, it wouldn’t run on YOUR computer. It would run on an Apple computer. If they were to try to enter the commodity PC market, they would be annihilated by Microsoft in instants. No, even Mac OS X for x86 would require some kind of Apple-specific BIOS or ROM so that it could not be run on commodity PCs and could not be easily pirated.
Even if Apple were to move to an x86 processor, it wouldn’t run on YOUR computer. It would run on an Apple computer. If they were to try to enter the commodity PC market, they would be annihilated by Microsoft in instants. No, even Mac OS X for x86 would require some kind of Apple-specific BIOS or ROM so that it could not be run on commodity PCs and could not be easily pirated.
Either that or they would have to pump up the price of the software majorly so that there a decent profit to pay for R&D as their computer side currently subsidises the operating system development.
It would be the eqivilant of a person saying that Microsoft should opensource their Office and Windows line up. Windows and Office subsidise other parts of Microsoft.
People forget the internal politics of what happens in a large corporation. One part makes a profit to subsidise the another part which enables the first part to exist.
As for G5 equipped Macs, IMHO, you’ll most likely see them once the PowerMac line up is refreshed, however, if it were me I would remove the 1.6 PowerMac, bring down the 1.8Ghz to the 1.6 pricing, and use the 1.6Ghz processor for the low end desktop range.
Even if Apple were to move to an x86 processor, it wouldn’t run on YOUR computer. It would run on an Apple computer. If they were to try to enter the commodity PC market, they would be annihilated by Microsoft in instants. No, even Mac OS X for x86 would require some kind of Apple-specific BIOS or ROM so that it could not be run on commodity PCs and could not be easily pirated.
Either that or they would have to pump up the price of the software majorly so that there a decent profit to pay for R&D as their computer side currently subsidises the operating system development.
It would be the eqivilant of a person saying that Microsoft should opensource their Office and Windows line up. Windows and Office subsidise other parts of Microsoft.
People forget the internal politics of what happens in a large corporation. One part makes a profit to subsidise the another part which enables the first part to exist.
As for G5 equipped Macs, IMHO, you’ll most likely see them once the PowerMac line up is refreshed, however, if it were me I would remove the 1.6 PowerMac, bring down the 1.8Ghz to the 1.6 pricing, and use the 1.6Ghz processor for the low end desktop range.
Apple’s hardware sales subsidize all the rest of their operations. It’s hard to deny that they make a pretty good profit on their hardware. It is, however, this profit that allows them to do things like develop OSes with no copy protection and to run the iTMS for no profit. Apple does a lot of things for free or no profit, but the one thing they can’t do is destroy their hardware sales.
then there is the HUGE problem with Longhorn not running on older hardware and the fact that most windows users run windows 98, so app writers will have to some how figure out how to support the people who are not upgrading because they like the machine they have or they do not want to change OSs since it is working fine for them.
What makes you think Longhorn won’t run on older hardware ?
Apple went through the same problem, but at least with apple. the OS could run on old hardware and run pretty well so almost all OS 9 users could upgrade their machines to OS X.
You must have been using different Macs and OS X to me. OS X was not only atrociously slow on older (and newer, for that matter) hardware but had poor compatibility with it to boot.
“That is what they do now using WOW (Windows on Windows) which enables one to run crappy win16 and non-pure win32 applications on Windows XP.
Sure, it isn’t an eligant solution but it kind of works š “
I haven’t gotten it to work reight yet.
“Microsoft won’t have a problem if they team up with OEM vendors and offer and incentive for users to upgrade, something like, “buy a PC with longhorn pre-installed and get Office Standard for only $20!”.
Microsoft needs to break compatibility BUT they’re unwilling to part with money to “subsidise” the initial longhorn adotion. If they can say to vendors, “here is some money, produce a pure longhorn version of your application” then subsidise a competitive upgrade scheme, “upgrade from win32 version to longhorn for only $20!”.
Microsoft needs to create that base for customers to move foward. Most users aren’t worried about pissant shareware rubbish but if they can’t get the applications they want on the new platform, they’ll stick with “what works” for them.”
Hmmm…kinda like with what I was offered two years ago when I bought my Dell (er Dull).
“How many Win98 users will there be in 2006? How many of those users will hold out for whatever is after Longhorn? Will computer techs remember how Win98 works?”
I still see Win95 around too. I seem to recall hearing that by Win2k nobody would supposely use Win95/Win98…yet hear we are. Next came the claims of XP…XP would supposedly be what would put Win9x out of its misery yet I still see hardware and software being manufactured fro Win9x and a whole lot of people out there who don’t feel like changing their OS or upgrading their systems.
Since 2006 is only a couple of years away I’d have to say that it is likely that there will still be loads of Win98 users who don’t feel like changing their computers.
“Most people are probably like my dad – gonna stick with his P3-600 for a couple more years until Longhorn is released. Until then, there’s really no reason to upgrade. But I doubt most people would stick with Windows 98 beyond 2006 and Longhorn, cuz you know the windows updates will have stopped by then.”
Hmmm…I didn’t think many people were using Windows Update. Remember MSBlaster?!
I couldn’t even begin to tell you how many people I have had to show how to use Windows Update.
I really don’t think an end to updates will mean anythig=ng…most people (out of those who even check for updates) will probably just think that it means that there is no reason to update their system.
That is what they do now using WOW (Windows on Windows) which enables one to run crappy win16 and non-pure win32 applications on Windows XP.
Sure, it isn’t an eligant solution but it kind of works š
What’s inelegant about it ?
I like how the M$ folks dismiss people that use other OS’ and say their on the edge/fringe.
Much like the people who use non-Microsoft OSes dismiss anyone who uses Windows as “sheep”, no ?
With the new advent of technologies that will be used in WinFX and the source of applications one would need to consider the support and overall throughput of the new OS. It would achieve higher ROI than lets say a solution of apple servers. With the cost of technology it is better to stay with a operating system such as Win Server 2003 with a track record which is proven.
It would achieve higher ROI than lets say a solution of apple servers. With the cost of technology it is better to stay with a operating system such as Win Server 2003 with a track record which is proven.
<p>Ah, I wouldn’t say Windowss 200*3* has more of a proven track record than OS X.
<p>In that context I’d say the primary reasons you’d choose a Windows+Intel platform over OS X – apart from the obligatory “is the software available” – would be hardware cost and/or capabilities. At the end of the day, intel hardware is cheaper and has a more powerful top end.
then there is the HUGE problem with Longhorn not running on older hardware and the fact that most windows users run windows 98, so app writers will have to some how figure out how to support the people who are not upgrading because they like the machine they have or they do not want to change OSs since it is working fine for them.
Most new operating systems have issues with older hardware. Its kinda how it goes for the whole industry
The software interfacing isn’t that big of a deal. Most of the enhancements a new version of a program running on Longhorn will bring are going to come from the new UI in longhorn. Many developers may not update older versions that run on 9X while offering newer versions for longhorn.
Also the current Windows UI is fully supported in .net. Longhorn’s new UI is .net. Its just the next evolution of WinForms. The backwards compatability is already there via the .net framework for making the UI transition.
“That is what they do now using WOW (Windows on Windows) which enables one to run crappy win16 and non-pure win32 applications on Windows XP.
Sure, it isn’t an eligant solution but it kind of works š ”
I haven’t gotten it to work reight yet.
WoW has been around for a _long_ time – since 1993 – it’s well proven technology. What Win16 or Win32s apps you trying to run that don’t work ?
With the talk of having OS X on Intel I was thinking what is keeping microsoft from making a windows version for the G5 so apple users could partition their apple drives and dual boot os x and windows? They have windows for PPC – I believe …..I think it would be very easy.
– JVM
Only windows NT ran on PPC, microsoft won’t prot to PPC, just like Apple won’t go to x86 it provides no benifits and all software for it has to be recompiled for the new architecture.
Windows NT wasw actually ported to PPC, Alpha, X86, and I think there was another. No other version of windows has been ported to something like that. Windows XP 64 bit comes close but those are still x86 at there hearts(??)
Windows NT wasw actually ported to PPC, Alpha, X86, and I think there was another.
MIPS. Oh, and the Intel i860, which was the CPU initial (and quite a lot of ongoing) development was done on.
I seem to vaguely recall it was actually some hardware feature missing from MIPS that was responsible for the algorithms that caused NT’s pre-Win2k poor VM performance – can’t remember where I read that though.
No other version of windows has been ported to something like that. Windows XP 64 bit comes close but those are still x86 at there hearts(??)
Windows XP *is* Windows NT (version 5.1, Win2k being NT 5.0). Win2k was “officially” on Alpha all the way up to RC2 and NT4 was rumored to have been ported internally to SPARC and PA-RISC (highly likely) in addition to the four platforms supported by the released version.
Similarly, internal ports to multiple architectures almost certainly exist for the current version of NT – XP – just like internal ports of OS X to different platforms – at least x86 – are kept in sync with the released PPC version.
Incidentally, NT has _never_ been “x86 at its heart”. Initial development was done on Intel’s i860 RISC CPU and once it became usable (the NT team was famous for “eating their own dog food”), it was always developed on multiple platforms so platform dependencies didn’t creep in.
With the talk of having OS X on Intel I was thinking what is keeping microsoft from making a windows version for the G5 so apple users could partition their apple drives and dual boot os x and windows? They have windows for PPC – I believe …..I think it would be very easy.
It would be easy, but it would also be pointless (heck, downright silly from a business perspective). PPC is/was a hell of a lot less compelling to Microsoft than Intel would have been to Apple.
all through this thread you have shown your ignorance when it comes to hardware over and over.
stop being an X86 fanboy….PPC is a great architecture. the G5 is barely slower than the Opteron and beats all other X86 chips. the fact that VT bought the G5s for their cluster rather than an opteron solution shows you that the G5 machines were cheaper. maybe next year opterons will come down in price, but who knows….the fact is that I could trip over myself to buy a PPC 970 chip from IBM and build my own box with it in there and run Linux or Windows on it because the chip is better than X86 chips out there.
Running Windows on a PPC970? Uh uh… curious to see that
IBM sell Power4 who has some little differences with a PPC970 too – and appart from some OS like Linux, you’re not going to use Windows nor MacOS X on it
Don’t compare apple and pear… bad
the fact that VT bought the G5s for their cluster rather than an opteron solution shows you that the G5 machines were cheaper.
One wonders about how this could have been so. Doesn’t sound like IBM or HP tried very hard. Ah, well, more fool them.
Also, given they’re going to have to run any important jobs _at least_ twice to account for possible memory errors, I’m wondering how well that cost/power ratio is going to work out with real-life usage. The fastest computer in the world isn’t much good if it only gives you the right answer 4 out of 5 times.
I’m not sure how they plan to “move to ECC systems in the future” either – I don’t think ECC RAM support is something that can be retrofitted.
They’ve got an on-the-fly error correction system in place. That’s why their 10 teraflops number is a good deal off the peak of 17 teraflops. The delta is larger than most other machines on the top 10. The NEC earth simulator, for example, has a sustained performance of 80% of its peak.
Also, they can upgrade to ECC in the future. The machine supports it — they just cheaped-out on the RAM they bought.
By peragrin (IP: —.nrp3.roc.ny.frontiernet.net) – Posted on 2003-11-08 13:03:21
“…microsoft won’t prot to PPC, just like Apple won’t go to x86 it provides no benifits and all software for it has to be recompiled for the new architecture.”
This is where .NET and Longhorn’s managed APIs come in. Managed apps made with the next .NET release (and later with LH’s version) will be portable to different CPU architectures, including 32 and 64-bit CPUs. The code can compile at install time or at runtime, taking advantage of the available hardware on the user’s system.
Also, with WoW64, 32-bit x86 code can run emulated on platforms that don’t have native compatability.
all through this thread you have shown your ignorance when it comes to hardware over and over.
stop being an X86 fanboy….PPC is a great architecture. the G5 is barely slower than the Opteron and beats all other X86 chips. the fact that VT bought the G5s for their cluster rather than an opteron solution shows you that the G5 machines were cheaper. maybe next year opterons will come down in price, but who knows….the fact is that I could trip over myself to buy a PPC 970 chip from IBM and build my own box with it in there and run Linux or Windows on it because the chip is better than X86 chips out there.
How did you come up with that theory?
G-5 is slowest of all chips today. Ever heard of Athlon FX-51 or P4EE. What makes believe that PPC 970 is better?
You have any solid evidence to support your theory?
All of the latest benchmarks that I have seen show that
Athlon FX-51 and P4EE are much faster in all but Photoshop test.
all through this thread you have shown your ignorance when it comes to hardware over and over.
stop being an X86 fanboy….PPC is a great architecture. the G5 is barely slower than the Opteron and beats all other X86 chips. the fact that VT bought the G5s for their cluster rather than an opteron solution shows you that the G5 machines were cheaper. maybe next year opterons will come down in price, but who knows….the fact is that I could trip over myself to buy a PPC 970 chip from IBM and build my own box with it in there and run Linux or Windows on it because the chip is better than X86 chips out there.
How did you come up with that theory?
G-5 is slowest of all chips today. Ever heard of Athlon FX-51 or P4EE. What makes believe that PPC 970 is better?
You have any solid evidence to support your theory?
All of the latest benchmarks that I have seen show that
Athlon FX-51 and P4EE are much faster in all but Photoshop test.
LOL!
Even Apple’s 2-GHz dual-CPU G5 unit had a hard time keeping up with a single-chip FX-51 PC in most tests.
http://www.pcworld.com/news/article/0,aid,112749,pg,7,00.asp
debman: stop being an X86 fanboy….PPC is a great architecture. the G5 is barely slower than the Opteron and beats all other X86 chips. the fact that VT bought the G5s for their cluster rather than an opteron solution shows you that the G5 machines were cheaper. maybe next year opterons will come down in price, but who knows….the fact is that I could trip over myself to buy a PPC 970 chip from IBM and build my own box with it in there and run Linux or Windows on it because the chip is better than X86 chips out there.
Uhmm, business-wise, this doesn’t matter at all. In fact, the G3 was cheaper thanm Pentium III, it didn’t matter then, it certainly doesn’t matter now. When Microsoft makes an OS, a large chunk of money would come from OEM deals. The last I check, Apple isn’t that interested in that. Why should they?
Another large chunk of their profits come from corporate sales. And here’s something that’s going to suprise you: x86 is dominant in the corporate market. If there’s any G5s lying about, the corporation wouldn’t have any reason to install Windows on it, they are probably used by the PR department or the graphics department or whatnot.
Then they also earn a significant amount of money via retail sales, but this hardly compares at all with corporate sales or OEM deals. And even so, Mac users interested in buying Windows would have still done so (One word: VirtualPC), there’s no profit is making a full port.
Oh BTW, I do agree that G5s are better than Opteron, on many different scales. If I was selling something like clusters and supercomputers, I would have definately picked G5 over Opteron. However on the desktop scale, things are very different. There is only one desktop company selling G5s- Apple. And they aren’t exactly selling it dirt cheap. Certainly, IBM has a lot to gain if Microsoft ports Windows to G5 like they did to Itanium, but not a lot for Microsoft to gain.
—
And on the comparison between OS X and Longhorn, other than it would be revolutionary for both Mac and Windows users, the comparison stops there. Why? OS X is not the continuation of Mac OS code, rather the continuation of OpenStep code. Apple had to merge both products and make something out of it. Longhorn on the other hand is based on Windows NT code and would be more or less as significant as Windows 2000. There’s no code rewrite, nothing of that sort.
And there’s another difference, a important one. Since Longhorn is based on the previous Windows, there’s no features missing in Longhorn that was available in previous versions of Windows – a case not true for Mac OS X then and even now. That way, migration to Longhorn is much simpler than with Mac OS X.
And there’s also another difference. Longhorn’s new API is .NET. .NET already runs on Windows XP/2000, older versions of Windows can more or less run Longhorn apps. By comparison, Mac OS X had the new Cocoa, based on OpenStep’s APIs, and no Cocoa app can run on classic Mac OS. They have to use Carbon, which has many disadvantages (it’s almost like using Win32). And there’s also another difference: Cocoa programming may be very easy (in comparison with Carbon), but porting old Mac OS apps to Cocoa is tedious and hard.
Transition to Longhorn, therefore, it much much easier than Mac OS X. Oh, there’s another difference. Rhapsody was first built on x86, as was OpenStep, and had to be ported to PPC. There’s no such problem for the x86 version of Longhorn. That also means when Longhorn comes out, there’s no worries that CD-RWs and DVDs don’t work, and the likes.
The PC-Mag comparison is a bit tad unfair. The AlienWare computer used today is $3,210.00, about $200 more than the highest end PowerMac G5 at $2,999.00. And the PowerMac wasn’t customized with RAID, meaning slower hard disk operations, in comparison with most of the other x86 machines. The difference in price allows you upgrade the graphics card to Radeon 9800Pro, in comparison with GeForce FX 5950 on the AlienWare.
And there’s also a lot of factors to speed other than the processor itself. Apple makes its own chipset, AMD hardly does. And since there’s a lot of competing chipset makers, from VIA to AMD themselves to NVidia, there’s a lot of competition causing the chipset (and therefore the motherboard) to be better.
Of course, I would certainly prefer to have the Aurora over the PowerMac – I much prefer the Aurora looks, and plus I already have all my Windows software (don’t need to buy new Mac software).
OMG, I’m defending the Mac! I feel so… dirty….
Oh BTW, I do agree that G5s are better than Opteron, on many different scales. If I was selling something like clusters and supercomputers, I would have definately picked G5 over Opteron.
What exactly do you mean BETTER? Can you be more specific? What makes it a better choice? For you personnel information Opteron scales better then G5. G5 has no ECC support and it is slower the Opteron, So what makes G5 a better cluster?
I reread the benchmarks more closely, I realize all systems were equiped with the Radeon 9800 with 256mb, all except one (ABS).
Anonymous: Slower than Opteron? Slower is definately relative, I have yet to seen one benchmark that manage to control all the variables. And it also really depends on 1) What OS the system uses, 2) what chipset the system uses and 3) What’s the cluster for. For that reason, I can’t give you a definite specific answer.
But remember, OS X for server operations is shamefully slow (I don’t know about Panther, but that description fits Jaguar fine). But all Opteron vs. G5 benchmarks I’ve seen uses Mac OS X on the G5. If I was to sell machines built for clusters, I would at least at the barest minimum create my own distribution of Linux/*BSD.
That’s why I find both sides of the debate “Opteron is faster!” and “No, G5 is much faster!” rather, well, funny. No one can definately say which is faster, not at least with the kind of benchmarks we have now. Oh, BTW, I can’t definately say, but I think that G5s is cheaper than Opterons too, in bulk. But you and I wouldn’t know, they don’t exactly have a pricelist on IBM’s site.
Oh, if I were selling clusters for something like cancer research, I would choose neither G5 or Opteron.
<< MS will have just as hard a time, if not harder moving over to longhorn >>
No thats what PDC was for so that their partners could keep track of Microsfts development process. And it worked. All of Microsofts major partners already have Longhorn and are undoubtedly working on it. One thing I give MS is that when they plan a leap like this they pretty much cover their bases. You are free to disagree if you would like.
<< then there is the HUGE problem with Longhorn not running on older hardware and the fact that most windows users run windows 98, so app writers will have to some how figure out how to support the people who are not upgrading because they like the machine they have or they do not want to change OSs since it is working fine for them. >>
The use of XP is rising so Windows 98 is phasing out. Lonhorn in alpha stages runs pretty well on a 700 mhz machine and by 2006 I doubt you will see less than at least a 2ghz machine.
<< Apple went through the same problem, but at least with apple. the OS could run on old hardware and run pretty well so almost all OS 9 users could upgrade their machines to OS X. >>
Not really, While you cpuld upgrade to OS X on a beige G3 and first gen iMac, performance on those machines was terrible and while Apple still boasts that OS X is compatible with those machines, I doubt performance with Panther really makes a difference. At least MS with their min. system requirements will run decent
<< Oh BTW, I do agree that G5s are better than Opteron, on many different scales. If I was selling something like clusters and supercomputers, I would have definately picked G5 over Opteron. >>
I disagree, I find the Opteron to be a mch better chip and if you want to sell the G5 go right ahead, your sales will be nil because the G5 is falling into a niche market.
“I reread the benchmarks more closely, I realize all systems were equiped with the Radeon 9800 with 256mb, all except one (ABS).
Anonymous: Slower than Opteron? Slower is definately relative, I have yet to seen one benchmark that manage to control all the variables. And it also really depends on 1) What OS the system uses, 2) what chipset the system uses and 3) What’s the cluster for. For that reason, I can’t give you a definite specific answer”
HD RAID array or MORE Video Ram will not create a Higher FPS in the game. The extra Memory is needed when you play at extreme resolutions with AA and AF enabled and nothing more.
They were measuring the FPS in the game and not how fast the Game Loads HD TIME.
Every test that pops up on the net clearly shows that Opteron is the winner and you telling me that it is inconclusive.
Opteron is the best price/performance CPU at the moment for cluster, rendering rig or regular workstation. Just read this here http://www.theinquirer.net/?article=10785
(He adds: “Basically the opteron scales amazingly well. I was shocked to find a 2.6 GHz Single CPU HT to be faster then a dual CPU P4 3.0 Ghz.”
He concludes: “The slowest Dual Opteron is 1.57 times faster (wall clock time) then the fastest dual Pentium 4”. Āµ) I know they are not talking about G5 here, but at least it gives you an Idea about Opterons amazing scaling capability.
(He adds: “Basically the opteron scales amazingly well. I was shocked to find a 2.6 GHz Single CPU HT to be faster then a dual CPU P4 3.0 Ghz.”
He concludes: “The slowest Dual Opteron is 1.57 times faster (wall clock time) then the fastest dual Pentium 4”. Āµ) I know they are not talking about G5 here, but at least it gives you an Idea about Opterons amazing scaling capability.
I suspect you’ll find these results are being *strongly* influenced by the bus speeds, which is why the single CPU machine is faster than the dual. Find something which isn’t bottlenecked by the bus speed and the dual P4 is almost certainly going to be faster than the single.
Like many Mac heads who aren’t blindly faithful to Apple, I would be OK if new benchmarks actually do show that AMD’s new processors to be faster than the G5. But I still haven’t seen anything of the sort.
PC World was ridiculous to run these specific tests, as a whole whack of messages elsewhere on this list suggest. Word is a great program for the Mac, feature-wise, but speed has never been its forte. And Premiere doesn’t even run on OS X; it has to be run in emulation mode using Classic. So the only relevant tests that PC World ran were the Photoshop tests, and the G5 was competitive.
Think of all the real-world, cross-platform tests that PC World could have run… Acrobat, Illustrator, Lightwave, Mathematica, BLAST, After Effects… well, it’s a very long list. I’m not sure if PC World has any journalistic integrity, but it seems as if these tests were chosen solely to prevent the G5 from being competitive.
Maybe we should benchmark new Opteron-based computers using Filemaker Pro, iTunes, Logic and QuickTime… Those benchmarks would have as much credibility as the PC World benchmarks.
Truth to tell, so far the G5 is proving to be a world beater, and much faster chips are reported to be in the pipeline. I posted these benchmarks on anothe OS News page to illustrate my point…
<http://www.osnews.com/comment.php?news_id=5060&offset=75&rows=90>…
I thought WinFX was a technology company founded in Switzerland way back in 1996.
Lawsuit, or does Microsoft not care any more?
And of course the nice WindowsFX / WindowBlinds people may have something to say on the name as well.
First thing that comes to mind when I hear WinFX is something multi-media oriented, not a programming API.
Oh well, whatever, I will destroy M$ anyway. š
–The loon
Okay, on the intel side:
The Taultitan (or whatever) processors core was considerably faster than the pentium4. I had a Celeron-T clocked at 1.5GHz on S370 FC-PGA with a converter for FC-PGA2 compat, and then I upgraded to a 2GHz pentium4 and was sadly dissapointed.
So I moved to an 1800+ Athlon XP and clocked it to 2GHz (real) and fell in love. I then had a customer request a 3GHz pentium4, it, too, was slower than my Athlon XP, so I stayed put.
Eventually, I moved on to an Athlon XP Barton 2500+, but my old mobo’s BIOS didn’t get along too well, even with the latest BIOS updates, so I downgraded to the XP 2400 (a real 2GHz) and clocked it up 2.25 GHz for a little extra something over my 1800+ that clocked so amazingly well (was a furnace though, system runs cooler now).
I have decided to not even pay attention to intel. I couldn’t tell you anything about their latest offerings if it has even changed. But I can tell you that I can’t wait till I move myself and my platform to 64-bit. Everyone will know when I get my Athlon 64 up and going, because I’ll sundenly make a huge jump into REALLY LOW LEVEL programming to best take advantage of 64-bit computations.
And G5… I think it is a bit behind what I am running, which is sad.. RISC is suppose to be easier to scale than CISC, so what is the deal? Oh.. right… no one wants to make them anymore… except now maybe IBM so they can finally sell some of their stuff they made… hehe
–The loon
I am frankly tired of people pointing to this benchmark. It has been pointed out to no end that the bench marks were unfair. Lets disect the PCworld amateur benchmarks one more time.
From the little information on Adobes site it looks like Premier was never actually ported and supported on Mac OS X So it makes for a really bad cross platform benchmark.
Word once more is not a direct port and a rewrite for the MacOS X and has not had as many years of development as the windows version and It is no surprise that it performs better on windows than on the Mac just as Quicktime and Itunes perform better on the Mac than windows.
So that leaves us with only two real benchmarks photoshop and Quake III
The Alienware Athlon FX PC without the extra Graphics memory and raid performance poorly on almost every benchmark compared to the setup with RAID and the extra Graphics memory.
Apple Power Mac G5
Two 2-GHz PowerPC G5s
128MB
Standard
photoshop 18 51
Quake 294 207
Alienware Aurora
2.2-GHz Athlon 64 FX-51
128MB
Standard
Photoshop 21 60
Quake 335 257
So that leaves us with only one configration for both the G5 and the Athlon Box. Now if you look at the numbers the Athlon is not much faster than the G5 in photoshop and Quake. I confiugred a Alienware Aurora with the same config as the G5 at $3426. So the price for the G5 is less.
I really would like to see a single CPU 2GHz G5 benchmark done because thread migrations can wreak havoc on performance on dual CPU systems becuase of losing a “warm cache” on migrations.
Anyway even with the higher scores in quake the Athlon box is more expensive and will definately not perform well on workloads with many threads due to being a single cpu box. So the G5 is a better value than the alienware box as a general purpose machine. if you really want better FPS in quake or games spend the money on the Athlon box.
The Athlon has a 1MB cache and the G5 a 512K cache thus making the cache a big factor becuas a smaller cache with two cpus can degrade performance on certain workloads because of cache coherency protocols getting in the way.
The bottom line is higher benchmark numbers means nothing if my workloads are totally different. The G5 will be good for certain workloads and the Athlon for some making a blanket statment on the performance of a machine with skewed and imporperly done benhmarks is not wise. Performance benchmarking and tuning is an art and unfortuantely pcworld is in the league of a kid with a magna-doodle in that field, proved by the benchmarks they posted.
<< For a while I kinda wanted a pc, that’s why I have this Dell. But now, with OS X and the G5’s and G4’s being so cheap(ibook, powermac etc). Who cares about crappy windows. I just wanted one, because I listened to everyone rant about spead.
My experience with XP has been love/hate, and mostly hate. It’s blue screen on my once, got infected with a virus and a worm(lacks security), it’s slow, hangs up, and its retarted cartoonish looking compared to my old iMac running OS X .1
So screw lame windows………. >>
What I find funny is that fr every 1 mac user that says screw windows, you have 5 or 6 PC users who say screw macintosh. I have no love/hate relationship with any OS, I use what i have to use to get a job done. A computer is a machine, glass plastic wires and metal hardly aything to love and hate. I do have operating systems that I like to work with, and for work I like Linux for play I like Windowsfor tinkering I like the Mac, I have butchered and frankensteined waaay to many Macs. For each one i like several different things, Linux I like the most because it requires you to think, to excercise your mind. Windows, I like the GUI, the abundance of programs and games and the overall simplicity of Windows. The Mac, I like the cases, I personally find the Mac OS X GUI Horrid, but thats a personal opinion. I have stripped old dead G4 boxes and thrown in a PC motherboard and installed Windows and Linux. I personally dont like Benchmarks and I pay no attention to benchmarks or any benchmark comparisons because frankly I dont care what is faster, Is the opteron fast enough for me to get the job i want done, yes. Is the G5 fast enough for me to get the job I want done on a Mac? Yes. Benchmarks are nothing more than a silly marketing ploy and anyone who will base their computing buying habits on a benchmark test are just as silly as the companies that promote their product using benchmark studies. They are inaccurate and biased, not just 1 but all of them because everyone who does these benchmarks all have a computer preference and the tests all seem to favor the testers preference. I have yet to see an honest benchmark, when someone releases a benchmark sayint that the G5 is faster you have another that says the Athlon 64 is faster. I will never change my main platform because even though one may be faster than the other, you lose something else in the process.
At work we took a little vote and we all agreed, Longhorn looks nothing like the Mac OS, we all think it looks like a cross between QNX and the MorphOS
āI am frankly tired of people pointing to this benchmark. It has been pointed out to no end that the bench marks were unfair. Lets disect the PCworld amateur benchmarks one more time.ā
More like you tired of G5 loosing to High end PC box! Iām willing to bet that you would sing a totally different tune if the situation was reversed. Donāt blame the PC world for the results.
Blame Apple for making Claims that they canāt support! Who is to say that apple did a fair test when they published there benchmarks. Marketing BS is what it is. So then what makes you believe that G5 is the āworldās fastest machine.ā You also state that Apple is a better Value. Really? .And what way would that be? Enthusiast PC users do not buy a prebiult unit. āDude you are getting a Dellā simply Dous not ring the bell. We build machines that fit our needs. Sorry to say, but Apple duos not offer any Value to PC users in any way. We have more hardware and software to choose from. A total freedom. Something that you will never have!
Whoa, I know this is really going off topic and really do not wish to post off topic. But just to provide closure.
Apples G5 benchmarks were open to interpratation just as PCworlds are. However, Veritest was very thorough in thier report and thier benchmarks were very well done. Just look at their client list you will find intel, cisco and many others other than Apple. Veristest is one of the best independant testing labs.
Also, PCworld was talking about Alienware and Apple out of which Alienware is more expensive for a single CPU config comapred to Apples dual G5. Looks like the athalon FX-51 is an expensive chip with a MB of cache.
I never said anything about DIY PCs, why bring that up?
Anyway looking at longhorn. There was a little discussion here about longhorn running on older hardware. I seriously doubt it. The minumum Graphics card required is a 3D one with 64 MB memory, 128MB recommended. We know when microsoft recommends a config it is the bare minimum. My home built athlon box only has 32 MB on Radeon (I am not a gamer) I have no hope of running longhorn on it.
Well after reading about longhorn one thing is clear it is a new generation of windows. But most of the graphics interface enhancements like transparancy and special effects are already in Mac OS X. With Apple’s yearly realeases I expect more to come and with G5s now in the pipeline expect G5s to replace all the G4s in the power[Mac,Book] line by 2006. So the shoot out will be 10.5/ 10.6 against longhorn and it will be one to watch.
The whole “just burn” thing for the photos in the video. Oh looks like the burn to disc logos in MacOS X are a big hit in redmonds Human interface design team.
The Athlon FX lost all the Photoshop benchmarks versus the G5 in a PC World article.
http://www.pcworld.com/news/article/0,aid,112749,pg,8,00.asp
The Premiere and Microsoft Word benchs don’t even count. If you look at ars and slashdot forums, people agree that these apps are poorly tuned for the G5. PCWorld could have used Avid XPress but didn’t for some reason. In NLE FCP and Avid are the programs people use not Premiere.
“The Athlon FX lost all the Photoshop benchmarks versus the G5 in a PC World article.”
FX did, but not the Opterons.
http://www.pcworld.com/news/article/0,aid,112749,pg,8,00.asp
Can we please drop the PCworld benchmark. Looking at polywell’s web site the poly station uses a 10,000 RPM SATA drive. A second on the benchmark for a 50 MB image and 4 seconds on a 150 MB image is a very consistent difference.
A 7200 RPM drive’s max write would be in the ball park of 30 MB/sec with reads of 35-45MB/sec. A 10Krpm drive would easily do 45MB/sec+ writes and 45-70 MB/sec reads. It is very supicious that the opteron workstation saved a second on a 50MB file and 4 seconds on a 150 MB file.
So till PCworld puts out detailed specs for each of the test machines, just comparing processors with this benchmark/hack is unrealistic. The CPU is just one component of the system. The 50MB file image has to be read and written to disk as well and a 10K RPM disk will surely speed things up.
So it isn’t possible to say one way or the other that the opteron is faster than the G5 from PCWorld’s ridiculous benchmark.
Ridiculous! Why? Because G5 did not performed 2.2 times faster over all as Apple has claimed. What performance do you expect from G5?
However: “The world fastest Computer” is the most ridiculous statement that any one has ever made!
HD RAID or MORE VIDEO Ram will not crank out HIGHER FPS in Q3 game.
You can read this Article here. http://www.tomshardware.com/graphic/20030604/index.html
There is plenty more articles just like that on the net
If you wish to read them all let me know I will direct you to them, but I feel it is absolutely not necessary to do so. All of them will tell you the same thing. This one has the better Graph that is all.
āA 7200 RPM drive’s max write would be in the ball park of 30 MB/sec with reads of 35-45MB/sec. A 10Krpm drive would easily do 45MB/sec+ writes and 45-70 MB/sec reads. It is very supicious that the opteron workstation saved a second on a 50MB file and 4 seconds on a 150 MB file.
So till PCworld puts out detailed specs for each of the test machines, just comparing processors with this benchmark/hack is unrealistic. The CPU is just one component of the system. The 50MB file image has to be read and written to disk as well and a 10K RPM disk will surely speed things up.”
That is just too funny! Alienware Aurora 2.2-GHz Athlon 64 FX-51 Standard (NO RAID 126MB VIDEO RAM) did the same time in Photoshop using 50MB image vs. Alienware HD RAID 256MB VIDEO RAM Version. The 150MB image with RAID did only 2 seconds better. Did you actually read the article carefully before jumping the gun? So if the tests are ridiculous to you, then please explain to me why RAID did the same time as standard set-up? Also FX-51 with RAID did much slower then Opteron with standard. Where is your Logic? And besides the world fastest computer should have won hands down on all tests no matter what you throw at it.
We can go back and forth all night/day but this is going no where and it isn’t the right forum to discuss this benchmark.
However, I think the PCWorld benchmark is ridiculous not because the G5 didn’t farewell but because the benchmarks weren’t chosen well and detailed specs of the machines weren’t given as any good benchmark publishing house would. Even the price of each of the machines is left out.
I am not defending the G5 just claiming that PCWorlds benchmarks are not a good comaprison even by thier own standards they measure CPU performance with different system configurations example RAID and 10K RPM drives with applications not optimized for every platform in thier test matrix and use applications that rely on other subsytems to generate good numbers. Thus failing as a reliable indicators of CPU performance.
“So if the tests are ridiculous to you, then please explain to me why RAID did the same time as standard set-up? Also FX-51 with RAID did much slower then Opteron with standard”
First the point I was trying to make is that the opteron system had a faster drive giving it a distinct advantage over the other systems RAID or not. That explains why the Athlon FX-51 did poorly on the test compared to both the opteron.
Quite simple RAID is not always going to be higher performance than regular SATA. You might want to read up on RAID before just believing that RAID in all cases is FASTER than a single drive with a faster spindle and seek time.
This website has a very good description of RAID performance issues
http://www.pcguide.com/ref/hdd/perf/raid/concepts/perf-c.html
“A lot of people seem to think that implementing RAID 0 is an easy way to storage nirvana, but for “average use” it may not help performance nearly as much as you might think.” Quote from above website.
Unless I read the Benchmark wrong the Alienware box without RAID took 2 seconds less than the RAID config on the 150 MB file (less time is always better!!!). Thus proving your RAID is better theory wrong. It looks like the Alienware aruora’s RAID is either not configured for the workload or poor on write perfromance.
Please let’s just end this debate.
“We can go back and forth all night/day but this is going no where and it isn’t the right forum to discuss this benchmark. ”
I agree with you on that!
I do have SCSI RAID 0 at home. I know how it works.
I use it every day.
I do a lot of Modeling and Viewing HIGH RES movies.
On my Single HD SATA drive the same movies are simply unplayable.
Also, I do agree with you that PCWorld did a very crappy job, causing this very debate.
We just have to wait for some one else to do the MORE FAIR and Extensive test.
This Debate is over!
PS. You can get your self a Radion 9800 NP and modify it into the 9800 Pro via BIOS FLASH. I know you were saying that you are not gamer, but you never know.
When will it be available?