This editorial at Rush Magazing talksback to Apple’s request to hear from PC users. The author has a problem with the speed, price and “complete hardware” solution Apple is providing. On the pros, the auhor lists the way the machines look and the fact that underneath you are dealing with a BSD UNIX. On a similar editorial at ZDNews: “But as it turns out, our reviews team recently did a hands-on with the fastest Power Mac on the market, equipped with dual 1GHz PowerPC G4 processors. Though it’s obviously the fastest Power Mac yet, it’s interesting to look at what you get compared with a high-end PC. The system’s P133 SDRAM, 80GB hard drive, 64MB Nvidia GeForce4 MX video card, and not-so-SuperDrive all look pretty pedestrian next to the Dimension 8200 and Pavilion 950, yet the Power Mac costs $3,000 without a monitor.”
Why I Don’t Own A Mac – Editorial at Rush Magazine
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSAlert.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
It’s simple. I just plain don’t care for the company. I do think they have some great technology, but I just don’t want to support them at this time. If they release OS X for x86 I just might change my mind, but until then open source software may be way behind but it’s good enough for me.
This story “Apple could easily migrate to X86”, http://www.theinquirer.org/01020211.htm , sums it up nicely. There is NO technical reason that MacOSX couldn’t run on off the shelf Intel/AMD hardware. Apple is simply afraid of what might happen if they sold an IA32 version of MacOSX. Microsoft would have a run for it’s money and Apple would likely capture tens of millions of NEW customers. What will Apple do when the IA64 Itanium comes out or the AMD X86-64? Apple is missing out of a huge market share opportunity by sticking with it’s propiertary hardware only solutions that cost much more than the comodity pc hardware. Many people would prefer to use MacOSX in their busineses except for the high prices of the custom Apple hardware. An Intel/AMD port of MacOSX will succeed in the marketplace and it’s about time that Apple took a risk and made it so.
We are chewing the subject on n’ on
It’s worth chewing on since an IA32 version of MacOSX is worthwhile. Maybe one day Apple will wake up and realize the size of market share they can capture with a MacOSX IA32 version.
“There is NO technical reason that MacOSX couldn’t run on off the shelf Intel/AMD hardware.”
I have heard rumors that OSX is already running on x86 internally at Apple. I don’t know for sure if this is true, but it would not surprise me at all. After all, given that it is based on MACH and FreeBSD, and that Darwin is running on x86, it probably would not be a big step to make the entire OSX run on x86. In fact, it would probably mostly just involve backporting code that for the most part, was intended to run on x86 originally.
Just my opinion… We are going to see OSX on x86. It’s the next logical step after Darwin becomes stable on x86.
I think, that on that shitty x86 with ancient legacy rooted to deeply, OS X ‘ll have very awful user experience. Not its unix layer, but all above.
And TeraHertz-es of x86 processor clock cannot change this fact.
I think, that on that shitty x86 with ancient legacy rooted to deeply, OS X ‘ll have very awful user experience.
In fact OpenStep 4.2, a previous version of MacOSX, runs totally awesomely on Intel and AMD chips. We have it running on an Intel Celeron at 450 mhz and on a 1.2 ghz Athlon. The responsiveness is amazing. There is no valid technicial reason that MacOSX would run slower in any significant way. The only apps that would be slower would be legacy Macintosh apps that were not recompiled “FAT” for IA32. Read the article “Apple could easily migrate to X86”, http://www.theinquirer.org/01020211.htm. It goes into detail about this.
OpenStep, and thus MacOSX, were meant to run on these “ugly” IA32 chips. While the X86 architecture is far from ideal it is low cost and is the best selling processor for desktop and server computers. This is why it won in the marketplace. Long ago I avoided the Intel x86 due to it’s “ugly” design and as a result I learned the hard way that this “uglyness” is irrelevant to the buying masses. Apple has still not learned this lesson it would seem and it’s costing them massive market share and revenues they could gain from unsatisfied Microsoft users and UNIX friends.
And TeraHertz-es of x86 processor clock cannot change this fact.
In fact it does change it! The IA32, since the Intel 386 and Pentium designs, has a core set of 26 or so instructions that essentially make it a RISC chip. This means that the IA32 is not all that ugly after all. While IA32 isn’t a beautifull chip like the Motorola 68040, Alpha or IA64, it is the best selling desktop processor line. It also means that the super high clock speeds of the Intel and AMD processors make a huge difference to performance. It it highly visible and noticable to users, even people who don’t use computers much can tell the difference. Just compare a real running copy of OpenStep 4.2 on a 33 mhz 68040 v.s. a real running AMD Athlon 1.2 Ghz and you’ll see what I mean. There is a huge noticable difference in performance of Applications and the OpenStep system when it runs on the Intel Celeron at 450 Mhz v.s. the Athlon at 1.2 Ghz. I’ll take the Athlon or Intel any day. Mhz DOES MAKE A DIFFERENCE regarless of the myth’s spread by some.
It’s simply nonsense to say that clock speed does not matter. You’re just buying into myth spread by Motorola and Apple regarding the PowerPC processor performance conquering all. There are many factors that enable fast processors, Mhz/Ghz, pipeline length and design, motherboard chip sets and buses, cache size and speed, RAM technology and speed, and more…
Instruction sets are largely irrelevant to determining processor performance these days. It has yet to be seen if the IA64 instruction set will provide any serious advantage over other instruction sets. This is why the ugly nature of the Intel instruction set is mostly a mute point these days.
>>The IA32, since the Intel 386 and Pentium designs, has a core set of 26 or so instructions that essentially make it a RISC chip.<<
Ah, the usual story about x86 being so RISC like! It may mimic RISC in concept, but it still isn’t RISC no matter how you slice it!
>>This means that the IA32 is not all that ugly after all. While IA32 isn’t a beautifull chip like the Motorola 68040, Alpha or IA64, it is the best selling desktop processor line. It also means that the super high clock speeds of the Intel and AMD processors make a huge difference to performance.<<
Just like Ford Motor Company saying they have the best selling trucks, it doesn’t mean anything from a technological standpoint, it’s a numbers game and nothing more, so get over the bull$h!t!!!
>>It’s simply nonsense to say that clock speed does not matter. You’re just buying into myth spread by Motorola and Apple regarding the PowerPC processor performance conquering all.<<
No you’re just a fool buying into Intel’s marketing ploy of selling clockspeed?… but if you’re dumb enough to fall for that idiocy, then more power to ya!
>>Instruction sets are largely irrelevant to determining processor performance these days. It has yet to be seen if the IA64 instruction set will provide any serious advantage over other instruction sets. This is why the ugly nature of the Intel instruction set is mostly a mute point these days.<<
Oh boy do you have a lot to learn… but what can you tell someone who has no clue on CPU architecture?!
But I guess since Hyundai sells more cars than Mercedes, they obviously have a better product?… NOT!! Bottom line is that the best technological innovations are never the best volume ‘deal’ sellers, PERIOD!!!
>”ugly” IA32 chips
After doing a significant amount of ASM programming, I wouldn’t say it was ugly compared to RISC ASM programming
>>Just like Ford Motor Company saying they have the best selling trucks, it doesn’t mean anything from a technological standpoint, it’s a numbers game and nothing more, so get over the bull$h!t!!!
<<
People please stop with car comparisons, everyone is always using auto industry as a comparison or example. What is with people doing this. Guess my big beef is most of these are invalid. When ford says they have the number one selling truck. It is a simple impressive fact. No one is forced to by a ford, their truck didn’t come preloaded with ford, buyers went to the dealership and choose it. They have the number one selling truck because it simple is far better than the compitition. It’s not like microsoft having the market share, people didn’t always have a choice in that matter. Also technology is not a big thing people care about trucks, its how good they just plain are, power, quality…… though the ford probly is the most advanced. Basicly not all numbers are BS some even though they may seam like spin on stuff, actully is a real number that is impressive. I think tech/computer/software companies are also capable of good non BS numbers. This doesn’t mean all are though. Also, comparisons can be good but make good ones.
It also has excellent FAT binary support.
You must be a Ford guy he he
The OSX on x86 debate can go on and on until Apple finally ports OSX to x86 (this date can also include never).
5 years ago I would have never thought I would buy a Mac, I absolutely hated it. I disliked the proprietary design and my inability to build one myself.
3 weeks ago I bought an iMac G4. On the first day of owning it I felt a little regret of buying it. I had overpaid for something that performs worst than my Athlon 1 GHz which I built for HALF the cost of the iMac. And the internal modem of the iMac just plain sucks, I can never get a decent connection, and had to resort to turning my PC into a router (proxy server worked fine for web browsing but not good enough for POP email).
When I had decided to buy the iMac, I promised myself I wouldn’t have too high expectations for it. My sole purpose of buying a Mac was to get away from Windows. I had tried Linux, but it did not work out for my purposes. Now that I’m getting more comfortable using OSX and programming it, I guess I can say I’m happy I bought the iMac. But still, there is nothing in this world that can convince me that the Mac platform with its PowerPC G4 CPU is better than an Athlon. Before you throw AltiVec in my face, bare in mind that x86 CPUs can make up for performance in higher clock speeds and lower cost. People say that overall cost of ownership between Macs and PCs gives Macs an advantage that makes up for the high initial cost of aquisition. That’s not true for everyone. Many people build rock solid PCs for way below $1000 that can outperform even the dual G4 1GHz.
These are esoteric arguments that are meaningless to the user. Remember what kernel_joe said, "it’s dark inside the box." The main thing is to get quality motherboard, memory, and hard drive. One will get mind-blowing performance out of OS X with any of the latest technology (Athlon, Intel, or PPC). I resisted BeOS x86 for a year and when I finally got it running on that platform I was completely blown away.
I don’t need or want Apple’s help in building what I feel is a good quality computer. I’m not convinced I need their OS either, but having it available for x86 would help.
Any argument beyond Apple not wanting to suffer the pain of being compatible with all the different x86 hardware is just meaningless detail and has nothing to do with end user performance. For anyone to get their panties all in a bunch over which processor a computer has is just geeky squabble.
>>That’s not true for everyone. Many people build rock solid PCs for way below $1000 that can outperform even the dual G4 1GHz.<<
That is depending on what performance you are looking for?
I’ll give you a prime example at a current situation… I am now sitting here at work with my laptop sitting on a console with a Sun Ultra10 to my left with a Dual 350 MHz UltraSPARC II running Solaris 7 (and our specialized software)… to my right sits a Dual AMD Athlon 1.6 GHz machine running RedHat 7.2 running the same specialized software (we are testing PCs to see if they’re up to the task of missions critical ops).
Last evening I was having a conversation with my good friend and colleague Brian about the performance traits of the 2 machines… he said in his opinion that though the AMD machine (which has 4 times the clockspeed) can do some simple tasks faster, it doesn’t do the hardcore tasks any faster an in some situations it is 2 as slow, now this guy actually owns an AMD Athlon 1.4 GHz machine and admits it is not as promising as he thought it was going to be, he runs WinXP and dual boots to RH Linux as well! He has come to terms that clockspeed doesn’t necessarily mean more CPU power, it just means a faster CPU, which doesn’t always mean a more powerful CPU! Basically the point is that it’s not how fast oyu’re going to get there, but how you got there and the obstacles that were in the way of your journey!
I am still not convinced on the MHz Myth and until I am proven different, I still will see Macs as a more powerful solution for my needs. Sitting right behind in the back right corner is 3 PCs stacked up with the keyboard monitor toggle switch (CYBEX SwitchView) to interface with the 3 PCs… they are all Compaq DeskPros with either a Pentium II or III running at 400MHz with 128 MB of RAM, they are all running Windows NT 4.0 (Service Pack 5)… now we take a similarly equipped iMac G3 running at 400 MHz with 512 MB of RAM (had 128 MB of RAM when I initially bought it in 1999 and had Mac OS 9) Though now it has 512 of RAM it wasn’t upgraded till 3 months after I recveived my copy of Mac OS X (10.0) and I can admit that Mac OS X original release was less than stellar, but when you compare it (and especially Mac OS 9) that the Compaq DeskPros were true sub-performers… I am not kidding! The sad part is that these machines are only for checking email, creating Word and Excel documents and web browsing, that’s all they do… our Sun hardware/software does the rest (well except for Linux filling in some areas like Network Management and our 1 test workstation I quoted earlier)!
Now in "T Ly’s" case I probably could believe some of his story… Mac OS X (and Windows XP for that matter) is more CPU and Memory hungry than old Mac OS ‘Classic’ (pick you flavor) helk I have an old Mac SE/30 with System 7 that is fast for it’s time… it is almost up to speed to our PCs here at work, NO JOKE EITHER!!!
Are Macs the most powerful machines out there?… helk NO! When you work around some of the industry’s most expensive hardware types like Sun, etc… you appreciate why companies are willing to shell out for this kind of gear, this stuff just plain works and does the job it was asked to do 24/7!
I am currently eyeing a PowerMac for my next purchase so I have been keeping a close eye on various benchmark tests from various websites, and I’ll keep those commentaries in mind, but I’ll be testing the machines myself before making up my mind! Some people are convinced x86 is the way to go, while others believe PPC is the way to go, that’s reality and that is what makes the world go around
Which CPU is better is complete irrelevant to the end user of mass market software systems. What matters is number of customers.
<P>
This is a fact illustrated by the success of the Intel line of processors. Apple is missing out on a massive market share by restricting themselves and their customers with propiertary hardware solutions. There is no technicial reason why MacOSX will not work on IA32 hardware. Given our real world experience with OpenStep 4.2 running on the latest AMD and Intel chips it’s obvious that clock speed makes a significant difference. Esoteric features of processors and which one is better than another is irrelevant. In our experience OpenStep blows away Microsoft software on the same power of computer. Thus MacOSX should perform in a similar manner. OpenStep and MacOSX are less bloated then Microsoft’s products.
<P>
Oh boy do you have a lot to learn… but what can you tell someone who has no clue on CPU architecture?!
I am an assembly language programmer. I’ve written entire systems in assembly language. I do know processors: 8bit, 32 bit and 64 bit. I prefer clean instruction sets like 68K, Alpha over ugly designs like IA32 and SPARC. Certainly IA32 is not all that bad anymore since the 386, except for the limited registers and complex instructions. Intel would not have succeeded against RISC processors with a CISC instruction set if they hadn’t invented processor implementation technologies that match or blow away RISC implmentations. For sure RISC designs use less transistors but Intel’s and AMD’s designs use other radical approaches that keep them performance leaders.
<P>
Certainly AMD’s 64bit extensions are nice. The IA64 is potentially very nice but is complex due to it’s instruction bundles. It is an unproven processor and it has yet to be proven in production that it’s instruction set will make a difference. I think it will but it has yet to be PROVEN in real processors in the real market place against other processor technologies.
<P>
No you’re just a fool buying into Intel’s marketing ploy of selling clockspeed?… but if you’re dumb enough to fall for that idiocy, then more power to ya!
<P>
I’m not buying into Intel’s, AMD’s, Apple’s or your ploys regarding processors. It’s results that matter. It’s a simple fact that IA32 processors from Intel and AMD are the largest selling desktop and server processors. The majority of customers buy them. This doesn’t make them the best, it just makes them the most "popular". Even if they have an "ugly" architectural design.
<P>
It’s obvious when you run a copy of OpenStep 4.2 on real hardware that clock speed does matter! OpenStep 4.2 runs fantastic on a slow Intel Celeron 450 mhz and it runs blindingly fast on an AMD 1.2 ghz Athlon K7. Benchmark results to follow.
<P>
How does reporting these FACTS of real systems that I use imply that I buy into the "clock speed" ploy of Intel? Your assessment of me is nonsense. Clockspeed is just as important as pipeline design, instruction set design and other factors. There are no absolutes in the processor business as new chip implementations are being produced at a frantic pace. A speed upgrade to an "inferior" processor can make it much faster than "better designed" processors. Intel and AMD are prime examples of this situation where one leads for a while and then they switch places.
<P>
But all this is mute. MacOSX will run excellent on Intel and AMD hardware. Maybe faster then on Motorola PowerPC processors, maybe not. It’s irrelevant which is faster for most customers. What matter is how many customers Apple has. I’m not suggesting that Apple ditch their propiertary hardware at this time. I’m simply putting forth the proposition that Apple would be able to gain a significant market share of customers with Intel and AMD processors. I want Apple to succeed and their propieritary only hardware solutions is preventing many people and busineses from choosing their awesome software systems.
>>Bottom line is that the best technological innovations are never the best volume ‘deal’ sellers, PERIOD!!!
<P>
Obviously the “best technological innovations” and “best products” are important to you. They are to me as well. I don’t suppor the idea that because something sells more that it’s better. I never wrote that.
<P>
I support the idea that Apple would sell millions and potentially tens of millions of NEW customers with Intel and AMD hardware versions of MacOSX. I have experience with OpenStep to know that it will succeed from a technicial standpoint. As such there is no reason that MacOSX wouldn’t also succeed. Most classic Mac Apps could be “recompiled” FAT for IA32 and an emulator could be provided as Apple did before for 68K.
<P>
The risk is that people would stop buying Apple’s overpriced hardware. Maybe this is why Steve Job’s and Apple have hesitated. Maybe they are waiting for the right moment to join the battle in the AMD/Intel world.
As promised here are the bench marks of REAL working OpenStep 4.2 systems that MacOSX is based upon. We ran these tests today on our real hardware that we use in our production environment. Our OpenStep boxes are dedicated systems as they are used quite a bit still.
First are the benchmarks for the original NeXT hardware followed by an Intel Celeron and an AMD Athlon. All of these systems are running OpenStep 4.2 which was the last version widely available from NeXT. We used NXBench, NWBench and DrivePerformance apps (search google. Maybe someone has these recompiled for MacOSX and would like to post their performance numbers for various MacOSX hardware?
Original Next Hardware – 33 Mhz color turbo slab
Compile 13.32
Webster 54.54
Graphics V/V 19.84 D/V 13.13
Drive Performance 1.73
NXFactor Graphics 0.63.
OpenStep 4.2 running on Intel Celeron at 450 Mhz with EIDE harddrives.
Graphics D/V 55.22, V/V 4.27
Webster 10.239
Compile 6.07
Drive Performance 15.60
NXFactor Graphics 5.87.
OpenStep 4.2 on 1.2 Ghz Athlon with SCSI-80 drives.
Graphics V/V 19.74 D/V 3.16.
Webster 7.65.
Compile 1.74
Drive perfromance 54
NXFactor Graphics 8.51.
The expensive NeXT computer of nine years ago has awesome software and regarless of the new processor architecture you get a massive performance increase. Yet it’s now running on commodity hardware that anyone can buy.
There is also a huge difference in performance between the slow 450mhz Celeron and a 1.2 Ghz AMD Althlon.
OpenStep 4.2 is from 1994 and hasn’t been optimized for modern hardware. The Pentium 90 range of cpus were the first cpus that OpenStep was run on. MacOSX will have been optimized for modern processors and should run even better than OpenStep on Intel/AMD hardware.
An obsolete Athlon 1.2 is bottom line now and not even made anymore. Now the minimum is an Athlon XP 1500. Why do you need the very best when the worst of current CPU’s (obsolete 1.2 ghz Athlon) have incrediblely desirable performance.
These numbers demonstrate what the old NeXT software can do on new hardware without any optimizations for the hardware. Just imagine what MacOSX will run like! Especially when using other components such as a Athlon XP 2000+ CPU with a motherboard running 333 mhz DDR RAM compared with PC133 mhz RAM.
The usability of the OpenStep windowing system and applications with an Athlon 1.2 ghz is so much faster and smoother than using Microsoft’s bloated 2000/XP on the same hardware. NeXT got it right over ten years ago and Microsoft still hasn’t caught up.
The real myth that Apple SEEMS to need to promote is that you need PowerPC processors to run MacOSX with awesome performance. OpenStep for Intel/AMD proves this myth wrong.
Is Apple running scared from fear that they won’t be able to compete with Intel/AMD hardware vendors should they bring out an Intel/AMD version of MacOSX? Are they afraid that they will lose sales of their propiertatry PowerPC hardware if they should offer Intel/AMD hardware or simply offer the MacOSX as a software only package?
The future is bright for Apple if they seize the day and take advantage of their lead before Microsoft catches up by sheer brute force. To infinity and beyond!
It was so long ago that it’s hard to remember that OpenStep actually ran first on a lowly 486 at 33 to 66 mhz with 16 to 32 megs of very slow RAM! Now it runs on the fastest x86 compatible chips without any modificaitions. This is a testiment to the skills of the NeXT staff. I’m glad that NeXT took over Apple! To infinity and beyond!
First I want to apologize for my immature assumptions and/or analysis of your knowledge and experience. That is something I usually don’t condone in that type of behavior, but seems lately that Apple has been getting the usual prejudice bad press from folks that just don’t get it, and probably never will!
>>Is Apple running scared from fear that they won’t be able to compete with Intel/AMD hardware vendors should they bring out an Intel/AMD version of MacOSX? Are they afraid that they will lose sales of their propiertatry PowerPC hardware if they should offer Intel/AMD hardware or simply offer the MacOSX as a software only package?<<
The only problem I see in this is that most of the Mac user base isn’t so Intel/AMD eccentric, well I think AMD might be a sort of link… I noticed that both Mac and PC users are sympathetic to AMDs offerings and that both sides use AMD as a bridge to be hardware agnostic, where true Intel or PPC (Motorola/IBM) hardcore types are not as flexible or as forgiving to the other. The other problem is that Apple would lose that special Branding awareness and would just be considered as another PC maker if they went that route and would probably destroy their Mac user base, which is the most dedicated user base to date, well other than the Amiga user base… and speaking of the Amiga, it is going to be going the PPC direction, which might spark enough interest in the PPC techno that it could give more excitement with the Mac and Amiga on one side of the fence and AMD/Intel PCs on the other!
Thanks for the info posted above:-)
How does one go about getting OpenStep for x86? I’m curious and like tinkering with operating systems. Plus I have a bunch of old PCs and would like to do something other than run Windows 98 on them.
I looked into GNUStep but their website states that GNUStep is not an operating system but an environment. That left me confused, but I did learn that OpenStep was originally a commercial operating system. Since it’s been so long, I wonder if they’re giving it away for free now, I don’t expect the “open” in OpenStep to imply that it’s open source.
Your best bet would be to find someone to buy a used copy from. Ask in the NeXTStep/OpenStep news groups.
You might want to buy a used NeXT computer which would entitle you a legitimate license.
True, they are a bit similar but Apple added a LOT of fat. The new Macs are pretty fast but still lag heavily in many areas. The UI responsiveness is just not there (yet).
Still, OpenStep for x86 kicks major booty speed-wise, given the features.
True, it’s based on not the most modern kernel, the filesystem doesn’t support journaling or softupdates etc. etc. but it’s still very nice and gets my vote for best UI of all time.
Ah, memories…