The Wall Street Journal is reporting that Apple and Intel executives have been in discussions about Apple using Intel chips in its hardware. Apple’s only official response at this point is that it’s “rumor and speculation,” but the Journal implies that it’s nearly a done deal. No word on whether Apple would use Intel chips across its product line or just in one or two products, like the Xserve. And this could all just be maneuvering to extract a better deal out of IBM. Losing Apple would be a major blow to IBM’s chip business.
I got it, Microsoft is using IBM chips for their X-box, does it not stand to reason that Apple is making their own console maybe even a Y-Box and using Intel chips. The OS will be a stripped down version of OS-X running x86-64 just like MS is doing with Windows on Power 5 or what ever the hell X-Box 2 is running on.
Oh bugger it, WHO CARES!
2.7 ghz G5 isn’t to shabby, especially compared to intels products which are fireballs at those speeds.
There is no such thing as a 2.7GHz PPC970. The fastest 970FX is a 2.3GHz part. The 2.7GHz PowerMac uses hand-picked 2.3GHz chips that are water-cooled and overclocked to 2.7GHz. It’s really no different than the “GeForce 6800 OC” (pre overclocked) parts you see video card makers sometimes selling.
Just use your head. If there was a 2.7GHz PPC970FX, why has XServe been stuck at 2.3GHz for so long?
Let me clarify myself. Fundementally, there is nothing wrong with cherry picking samples and overclocking them. After all, CPUs don’t have a set clock, they just have a target clock and a maximum that they’ll run at (which varies from sample to sample). If a given 2.3Ghz part can run at 2.7GHz, without jacking up Vdd too much or exceeding temperature limits, then it’ll last just as long as if it had been rated to run at 2.7GHz. The reason I bring up the comparison, however, is that if you allow watercooling for the G5, you have to extend the same courtesy to x86 competitors. A good Athlon64 sample with a watercooling system can easily hit 3+ GHz. Now, you’re back to arguing the same thing, except now its 3.0 Ghz vs 2.7 Ghz instead of 2.6GHz vs 2.3Ghz.
Pff…forget x86 or PPC.
I WANT CELL!
//PowerMac
Kelson Rocks !
I haven’t really given this subject too much thought, but to me it seems that if Apple was going to completly swtich to Intel processors the ideal time would have been when OSX was first released. It doesn’t seem practical to switch now, considering how long it has taken people to transition from Classic to OSX. Forcing venders to rewrite their software, again, would be a stupid move.
Sure it’s possible, but I doubt it will happen any time soon.
1. I think this is probably about things like the iPod, and other small devices Apple might have planned for the future (new PDA, perhaps?).
2. OSX on standard 32-bit x86 is simply not going to happen.
3. Intel’s NIH attitude has calmed down somewhat past five years or so. They already license ARM and x86-64 (_from_ AMD). There is no real reason that Intel couldn’t license the PPC instruction set. In fact, I think this is the second real thing that could happen. IBM and Freescale haven’t made much progress to a new low-power PPC chip for use in notebooks, eMacs, and Minis. There is only so far that they can overclock the G4 (yes, that’s what they’ve essentially been doing), perhaps up to 2GHz tops. Freescale is going to come out with multi-core versions of the G4, which might work for the mini and eMac, but would probably draw too much power to be practical in a notebook. IBM hasn’t delivered a G4-class notebook/embedded chip. AMD and Samsung have both licensed PPC, but they also haven’t delivered a new mobile processor. Let’s face it — the Pentium-M is a pretty impressive chip. If Intel can put that research to work in a 64-bit PPC package, Apple would buy a ton of them.
4. This could also be a last gasp for the Itanic. The only company selling them in any significant numbers is SGI, and they’re bleeding cash very quickly. If Apple could port OSX and XGrid to Itanic, you’d have a pretty compelling platform for high-end tasks. Emulation of PPC instructions on Itanic should impose less overhead than x86 emulation. The actual porting job shouldn’t be that difficult, as, from what I understand, IA64 is not all that much different than PA-RISC, for which there was a native NeXTstep port.
I don’t know about Dell latest offerings, but at one time they did have Pentium4s in their notebooks. Here’s what I mean when I talk about desktop replacements.
Oh, I’m not arguing that they never have used P4’s – just that at least one of their new desktop replacements is using a Pentium-M. The Inspiron 9300 and XPS Gen 2 both are – the XPS’s is at over 2GHz apparently so it’s hardly the P-M you see in most notebooks!
I’m curious where you can support this claim about the overclocking and such.
I really doubt apple needs to water cool things. Just like why they use a Gigantic heat sink on the Powermacs, they do the watercooling to keep the noise down. I doubt even with just fans it would be very loud, but apple is trying to keep their boxes really quiet after getting a bad wrap for the extremely loud G4 Powermacs.
And at anyrate, their is nothing stopping dell or whoever from having factory watercooling setups.
I don’t count overclocking for anything, unless it’s coming from the factory that way, which OEMs in PC land don’t do, unless they are really sketchy.
You can’t say Apple is overclocking the chips since them and IBM are developing them and know what they can do, and warranty them. Even in x86 land amd and intel hand pic the best cpus to ship out the door with the higher rated clock speeds. In the end, if apple ships it, and list it as X.X ghz, then thats what it is, and it’s not an overclock.
Good to go Steve.
the G5 is stuck at 2.3 GHz* for what seems like ages.
Err The latest G5 runs at 2.7GHz in the power mac G5 dual.You might want to check their website more often.
Apple would go to AMD, and have them take a AMD64 and start ripping anything that is x86 left in it away, have them add registers, and basically end up with a new chip, a pure or close to it risc chip, designed by AMD for Apple, they wouldn’t bother trying to have all the old legacy x86 stuff in there since they don’t need it.
Then they might as well stay with PPC, because if they did that they would still have a single supplier and small numbers, meaning high prices.
Although the cruft in the x86 instruction does cost a certain number of transistors, especially in the decoder stages, it matters less and less as the overall number of transistors increases. Furthermore x86-64 has 16 registers already (as does the ARM), anything more doesn’t make much difference.
Meanwhile, RISC architectures have the problem that the higher number of instructions they require combined with their fixed 32-bit instruction format wastes memory bandwidth and cache space. Hence ARM’s 16-bit-wide Thumb instruction format.
If Apple did switch to x86, Macs wouldn’t become standard PCs and MacOS wouldn’t run on any PC, in order to protect the hardware business and the Mac’s superior hardware/software integration.
It doesn’t seem practical to switch now, considering how long it has taken people to transition from Classic to OSX. Forcing venders to rewrite their software, again, would be a stupid move.
This switch would be much easier, because the API stays the same, so in many cases a recompile (and re-testing) would do. Only assembler code, mostly Altivec stuff, would have to be rewritten. Given that Altivec and SSE2 aren’t worlds apart from each other, that shouldn’t be too difficult either.
Various artstechnica review of OS X have pointed out how it seems to be prepared for such a switch.
By the way todays computer industry is going alot of companies making alot of mistakes lately. apple is probably talking about jobs exclusive restroom toilets automatic flushing hardware switching to intel chip.
Phil:
What’s to stop Intel from making Power processors? It’s an open architecture, is it not?
matt
why couldn’t this be about intel making PowerPC chips?
ROTFL. It’s not going to happen…
that rumor appera at least 1 time a year…
More often. Maybe because it’ll happen soon.
http://www.internet-nexus.com/default.htm
Don’t turn an Apple into a mainstream box!
They might be expensive, but at least they are ‘different’…if i wanted an Intel box, i’d buy Dell (or something).
The OS and the design ‘make’ an Apple for most part, but i’d hate to find ‘Intel Inside’
If Jobs’ for real on this one, i’d better hurry and buy a PPC equipped one before they sell out!
If they won’t do it soon it’ll be pointless later because the next release of Windows will be very good. Apple will loose that small market share they have now.
I and a lot of ppl won’t switch to mac hardware, i rather wait for Longhorn. hehe
If they won’t do it soon it’ll be pointless later because the next release of Windows will be very good
When will longhorn be released?Will it include enough features to justify an upgrade other than for renewed service cycle?
I personally think MacOSX on carefully selected (supported) X86(_64) hardware if you see it as an addition isn’t so far fetched.
If they won’t do it soon it’ll be pointless later because the next release of Windows will be very good.
Never mind the quality or not of the next Windows, there’s no chance Apple could try to compete on standard PCs.
First, they haven’t got the drivers or the resources to test OS X on the myriad configurations out there.
Second, MS wouldn’t provide Office for PC-OSX.
Third, they’d kill their hardware business, because how many people would still buy Macs if you could get an OSX-compatible box from Dell or some Chinese outfit for half the price?
No, the only half realistic scenario is a proprietary Mac that happens to have a x86 rather than a PPC processor.
PocketPod!!
Apple breaks backward compatibility every 7 years …
That has to be the saddest thing I’ve heard. You actually choose a system because it’s different? You have that strong need to feel different?
“To hell with efficiency, I wan’t something other people don’t have!”
I’m not a Mac/Win/Linux zealot, I just find that reason really ridiculous.
you are right about the RISC code issue needing more instructions, but the performance difference is that in CISC you need to decode the instruction into its component register operations which can take much longer than what the equivalent RISC code would do since a RISC operation is directly related to a single register operation.
so you have a need to conform?
so you have a need to conform?
Don’t think so,trying merely to conform is as meaningless as letting your OS opinion being formed solely by nonconformism.
is this at all relevant?
http://www.power.org
“Proposed Mission
Power.org’s mission is to develop, enable and promote Power Architecture technology as the preferred open standard hardware development platform for the electronics industry and to administer qualification programs that optimize interoperability and accelerate innovation for a positive user experience.”
“The foundation for Power Architecture was originally developed by IBM in the early 1980s and subsequently steadily improved through collaborative efforts with leading brands and developers from around the world. Initially used mostly by IBM, Motorola and Apple, Power Architecture™ technology”
IBM wants to become ARM but with a foundry and they are pushing that powerpc core into everything. That suggests that they are distracted from Apple but it also suggests apple could pick up a new foundry for the powerpc core and maybe even the G5.
AT the end of the day, it is more important for IBM to spread the powerpc than it is for them to fab apple’s chips. IBM has plenty of other customers.
If Apple jumped into the Intel arena, how would it maintain its exclusivity? The new x86 Macs would no doubt be equipped with a boot ROM or BIOS that would be required by Mac OS X to run. But how long would it take competitors to reverse engineer those chips so that Apple’s crown jewels would run on any PC box? A few hours? A day? Didn’t IBM try that sort of thing years ago, and how long did it take for PC clones to take over the market?” The above Quote was taken from the Mac NightOwl. Its not going happen for the reasons stated above. One day a week tops to reverse engineer the bios chip and let all cheap in quality Pcs run OSX. Again Won’t happen.
I’m curious where you can support this claim about the overclocking and such.
It’s speculative, but I’m hardly the only one who believes it. There is just too much evidence to support the idea. First, read this:
http://www-03.ibm.com/technology/power/newsletter/august2004/images…
Second, note that the article specified the maximum working frequency of the PPC 970FX, 2.5GHz at 1.3v. Also note, however, that they recommend against clocking the processor at 1.3v, because it halves the life of the processor. So the reliable maximum is 2.3GHz at 1.2v.
Third, note that if we assume Apple follows IBM’s recommendation of 1.2v, the power usage of the 2.5GHz G5 could be no more than 70 watts, a mere 5 watts more than the 2.3GHz G5. The power usage of the 2.7GHz G5 would be 76 watts, or 11 watts more than the 2.3GHz G5. If Apple can quietly and effectively cool the 2.3GHz G5s on aircooling there would be absolutely no reason to go to watercooling just to dissipate an extra 5-10 watts. It makes no sense to incur the extra risk, cost, and complexity when you could dissipate the extra heat by just turning the fan up a couple of hundred RPM (which would increase noise only imperceptibly). So its very difficult to believe that the 2.5GHz and 2.7GHz parts are running at 1.2V.
Now, if we assume that the 2.5 and 2.7GHz parts run at 1.3v, the picture changes quite drastically. bcause of leakage current, the power usage increases more than quadratically to 100 watts at 2.5GHz, and a projected 108 watts at 2.7 GHz. Jumping from 2.3 to 2.5 or 2.7 isn’t just 5-10 watts anymore, but 35-45 watts, or a 55-70% increase. Suddenly, watercooling looks very attractive. Indeed, reports are that even under watercooling, the temperature of the G5 PowerMac still gets rather high. That’s quite consistent with what you’d expect from a rather small radiator (like the one in the PowerMac), being asked to dissipate 220 watts of heat.
Fourth, note that the product lineup isn’t consistent with there being “real” 2.5 and 2.7GHz parts. If there really was a 2.5GHz G5, shouldn’t it have been bumped down to the midrange PowerMac model instead of being eliminated from the line entirely? After all, that’s what happened to the 2.0 when the 2.3 came out. If there was a real 2.5 or 2.7GHz part, shouldn’t the XServe line offer them? If there are real 2.5 and 2.7GHz parts, wouldn’t IBM’s own JS20 line offer them?
So basically, IBM says a 2.7GHz 970FX isn’t possible, the mathematics of the situation suggests that the 2.7GHz part doesn’t run at IBM’s recommended 1.2v, and the product lineup isn’t consistent with the existance of real 2.5 and 2.7GHz parts. Everything points to the 2.7GHz G5s being hand-picked, overvolted and overclocked 2.3GHz parts.
I doubt even with just fans it would be very loud, but apple is trying to keep their boxes really quiet after getting a bad wrap for the extremely loud G4 Powermacs.
You can quite easily keep a 70 watt part cool using just a large heatsink and a quiet fan. 69 watts is the rated thermal design power of the new Athlon 64’s, and if you take a big heatsink (Thermalright XP-90), and a quiet fan 92mm fan (eg: a Nexus), you’ll get good temperatures and won’t be able to hear the thing through your case. Cooling a 110 watt part is a different matter entirely, as anybody who has tried to quietly cool a Prescott can attest to.
And at anyrate, their is nothing stopping dell or whoever from having factory watercooling setups.
Well, yes there is. Apple has total control of the hardware. They can hand-pick choice G5 samples that can run at the required clockspeed, put them in a case specifically designed to handle them, and sell the few machines they can put together this way to their small niche of buyers. Dell can’t do any of these things, because of their size. Dell isn’t really the best comparison to Apple. A far better comparison are the small “boutique” computer shops, which will put together a watercooled, pre-overclocked computer for you.
In the end, if apple ships it, and list it as X.X ghz, then thats what it is, and it’s not an overclock.
It depends whether you’re trying to compare CPUs, or “products”. Now, this is at thread about CPUs. This whole time, people have been talking about the G5 or the PPC970, not G5 PowerMacs. If you’re comparing just the CPUs, then its only fair to compare them under the same conditions (namely, comparable cooling setups). The fact that AMD can’t easily ship 3GHz Athlons, because they can’t count on their OEMs all watercooling them, is really a tangential issue (a management issue), that is not interesting at the technical level.
Let me put it this way. If Alienware took choice Prescott samples, watercooled them, overclocked and overvolted them, and released them as 5GHz computers, would anybody say that “Intel has a 5GHz P4”?
Hmm, meant to link to the whole newsletter, not just the image:
http://www-03.ibm.com/technology/power/newsletter/august2004/articl…
Listen and listen to me good, you maggots:
THERE – WILL – NOT – BE – A – MAC – RUNNING – ON – X86!!
A code base like a song by a beat poet on a processor that’s kludged together like a russian car for people who think paying for software and hardware is a crime against humanity. It is not flippin’ happening [if it hadn’t been for the fact that I respect Eugenia, I’d have said ‘fucking’, don’t think I’m trying to spare YOUR feelings].
Steve is not going down the road that has proven disastrous for Apple in the past, with a business plan that he killed as soon as he came back to 1 Infinite Loop. Nah ah!
It is quite likely he’s going to use Intel chips to drive the next insanely great piece of hardware that the boys in the basement of the dungeon are working on, but it’s not going to be a Mac.
Besides: for all the sporting people who indicated the MHz gap: how long has it been since Intel tried to breach the 4 GHz barrier? Why are they not in PCs today? Is it because they are *ssssssss* a tad too hot? Do you think?
The present speed of processors is proving to be good enough to run most tasks on a Mac as fast as on a PC [and spare me the analysis, will ya?]. The thing that Apple direly needs to lure in more PFYs is: FAST graphics cards. Something that can stand the rigors of what is it that you guys do at LAN parties… Zoo Tycoon, whatever.
Good graphics cards to compete in the games market.
Microsoft is already developing games on the Mac, the Mac market might just as well benefit from their availability as well.
Better graphics cards, more hybrid games, whatever it is that Steve has under the hood and you can stick your x86 chips where your com port don’t shine, sugar!
… and Intel sells the best ARM-based processor series, more powerful than the PortalPlayer chip used today. The new video features in the last update of iTunes and what Jobs says today about “podcast” and “video store” makes easy to think that Intel Xscale processor series, it used in both PalmOS- and PocketPC-based PDAs, will be perfect for a PDA-like iPod. And the well known Apple Newton used Intel StrongARM processors.
A code base like a song by a beat poet on a processor that’s kludged together like a russian car for people who think paying for software and hardware is a crime against humanity.
A code base like a what? Do you know anything about the OS X codebase? The base system contains code from 4 OSs (Mach, 4.4BSD, FreeBSD, and NetBSD), written in 3 languages (C for the kernel, C++ for the IOKit, C and ObjC for the toolkits), with two standard toolkits.
You want a clean codebase, read the NetBSD source. That’s clean.
“If they won’t do it soon it’ll be pointless later because the next release of Windows will be very good ”
Very good eh? Very good as in only half as many Web browser exploits as in previous iterations? And maybe you will get WinFS and maybe not? And just maybe you will get to install Longhorn before your kids go to college? Well, okay.
Where Longhorn goes, there is one word that comes to mind:
bloatware
“It’s speculative, but I’m hardly the only one who believes it.”
So, if 60 percent everyone believes the world is flat, then it’s flat?
“Third, note that if we assume…”
“Now, if we assume….”
I really believe that you assume too much. The reality is that people who like intel ad amd, can go to google, do some searching and find articles that will say their processors are better. People who like PPC’s, can go to google, do some searching, and find articles that will say their processors are better.
Different sites will have different benchmarks,so, what do you do, “assume” one site is more valid than the other? Average out the benchmarks?
You really can’t compare a PPC with an x86, no matter what. They are that different. Again, can’t compare apples to oranges….dogs to cats…etc.
Congratulations taking “if we assume” completely out of context. “If we assume” is standard language when trying to prove something by negation. I wasn’t actually “assuming” anything — I was pointing out the ramifications of various assumptions, and to the extent that those hypothetical ramifications do not match what is actually observed, those assumptions must be incorrect, and other assumptions must be correct.
I didn’t google an article to find one that said “opterons are better”. Heck, I never said opterons are better. I dug up an IBM tech document that said that the PPC 970FX runs at a maximum of 2.3GHz at 1.2 volts, then warns you not to overvolt it to 1.3 volts. Call me crazy, but I’m liable to trust such a document.
As for “you can’t compare”, of course you can! You can comapre anything, given a proper metric. At the end of the day, x86 and PPC are processors. There are standard ways to compare processors (SPEC), and SPEC says that the G5 and the Opteron are comparable in performance. I trust those results, just as I trust ASTM results that compare the strength of steel and plastic, even though steel and plastic are very different materials.
What to do if you’re Apple?
The XBox 360 will be bigger than the Mac, it practically is a Mac. If it runs Windows (and it does) then you could run Windows on your Mac. MS with likely sell many more XBox 360s than Apple will sell in Macintoshes next year.
The software runs on a G5 with an ATI card. All the people who say the XBox 360 is not PPC are all wrong.
Now if you’re Apple, and people can run Windows on their Mac, possibly even Longhorn, what does that mean?
If you’re Apple, and the 9000 pound Gorilla is working with your chip vendor, and you will get squeezed out, what would you do? Apple has to hedge their bets. MS will get IBM’s best effort and sell 20 million XBox 360 systems. Apple will not get IBMs best effort, and they will suffer if they wait around too long.
XBox 360
The new x86 Macs would no doubt be equipped with a boot ROM or BIOS that would be required by Mac OS X to run. But how long would it take competitors to reverse engineer those chips so that Apple’s crown jewels would run on any PC box?
And why don’t they do that at the moment, i.e. build clone Macintoshes like Umax & others did in the 90s? Apart from the processor today’s Macs mostly use PC components already anyway.
Now if you’re Apple, and people can run Windows on their Mac, possibly even Longhorn, what does that mean?
Nothing. People buy Macs for their design, their integration, and their operating system. PPC-based Windows machines wouldn’t change any of that. In fact it would be good for Mac users, because they could run PPC-Windows programs in VirtualPC without performance-munching processor emulation.
That would be interesting if I could run Longhorn on my Mac without the aid of Virtual PC. Maybe this is why Microsoft is taking it’s sweet ol’ time on Longhorn. I would have to believe MS aquired VPC for more than the Mac plateform. It would be cool if Longhorn had a “virtual switch” or shift that could reconize a PPC.
The google part of my post wasn’t directed toward you, it was directed toward other posters, and I was too lazy to go back and get their names.
As for your assumptions,
“I was pointing out the ramifications of various assumptions, and to the extent that those hypothetical ramifications do not match what is actually observed, those assumptions must be incorrect, and other assumptions must be correct.”
Just because those assumptions are incorrect, does not make others correct.
“As for “you can’t compare”, of course you can! You can comapre anything, given a proper metric. At the end of the day, x86 and PPC are processors. There are standard ways to compare processors (SPEC), and SPEC says that the G5 and the Opteron are comparable in performance. I trust those results, just as I trust ASTM results that compare the strength of steel and plastic, even though steel and plastic are very different materials.”
You’re going to like this response. I am going to start by using your logic that both are processors. Ok, an apple and an orange are both fruits, they grow on trees, and they have seeds. Ok, does that mean they taste the same? Have the same texture? Have the same smell? Provide all the same vitamins, nutrients, and minerals? Have the same insects and animals feasting off of them? Can be grown in the same climates? That they are the same color?
So, i guess you can technically compare them, but, after comparing them you cannot accurately measure performance, because of all their differences.
Just because those assumptions are incorrect, does not make others correct.
There are only two mutually-exclusive assumptions here. If one is incorrect, the other must be true.
1) The 2.7GHz G5 runs at or less than IBM’s limit of 1.2v for the PPC970FX
2) The 2.7GHz G5 runs at greater than IBM’s limit of 1.2v.
Since the math suggests that the processor generates too much heat to run at 1.2v, it must run at some number that is higher than 1.2v. Since this is over IBM’s voltage guideline, this implies that the processor is overclocked.
As for the comparison: you have to define a metric. If you’re metric is weight, then sure, you can compare them. Even if your metric is something more complex, like “good for you”, you can still compare them! You just have to define the right metric (say, average increase in life expectancy in controlled groups consuming each product). The same is true of processors, if you define appropriate metrics (eg: which runs a C implementation of bzip/gcc/perl faster, which is basically what SPEC is), then you can compare them.
By desidaerius504 (IP: —.fuse.net) – Posted on 2005-05-24 19:59:10
Where Longhorn goes, there is one word that comes to mind:
bloatware
OK, it’s your opinion, but please don’t cry if the next Windows will kill Mac OS X.
Anyway… It’s ridiculous. Are you sure that the biggest software company can’t come up with something really good? Is it good to be a rabid fanboy?
Just because those assumptions are incorrect, does not make others correct.
Instead of general handwaving, how about engaging with Rayiner’s actual, well-reasoned arguments.
So, i guess you can technically compare them, but, after comparing them you cannot accurately measure performance, because of all their differences.
Well, Apple and IBM seem to think that you can compare performance, as they’re both members of the SPEC consortium.
If you’d said that performance didn’t really matter anymore for most applications, you’d at least have a reasonable argument, although plenty of people would beg to differ. But saying you can’t compare processor architectures as similar as that is just pathetic.
Projects like Debian and NetBSD demonstrate how interchangeable most processor architectures are these days, it mostly boils down to switching compiler backend and adapting the low-level parts of the OS kernel.
I think that nobody here actually knows the price Apple pays IBM for its CPUs. Unlike x86s, bare PPCs are not sold in computer stores. I’m convinced that more PPCs are used for Apple computers than for workstations and servers ( excluding the embedded market which Motorola/Freescale targets ). So, believe it, the PPC product line depends on Apple.
With some exageration, IBM Microelectronics is an ASIC manufacturer. ( Application Specific IC ) : Some are CPUs for Apple Computers, others are CPUs for game consoles, … The economics are quite different between ASICs and ordinary hardware, like a x86 or a RAM chip.
The ‘overclock or not overclock’ discussion about watercooled PPCs is a bit biased. When Intel builds a Pentium, it doesn’t know in which mainboard it will be used nor the ambiant temperature in the computer box. [IMHO] Apple engineers have a much wider knowledge of the PPCs hardware limits than Dell’s have of the x86s. The real world is analog. IBM can select better than average chips targeted for overclocking. ( As a example, I’ve once used 400MHz Motorola PPCs, screened for low thermal dissipation, remarketed by ATMEL as 300MHz chips for use in extended temperature range ( -40°C to +85°C ambiant ) airborne systems )
If Apple would switch to Intel, they would switch to Itanium and not x86s. Competing with PCs doesn’t make sense.
If IBM is unable to build more powerful PPCs, don’t expect Intel to do it better. IBM knows a great deal in CPU design. Maybe PPCs are, in the end, inferior to x86s, Maybe too few PPCs are produced to sustain pouring as much money as Intel and AMD for enhancing the engine, Maybe 3GHz chips for desktop computers are useless except for hardcore gamers which should rather buy game consoles. ARM was a low end RISC design ( no multiplication and no FPU… ) which was simpler to design, needed less transistors and was faster than the 80386 & 80486. There is currently no smart alternative to PPCs and x86. The only new design, Itanium, is _very_ complex and marginally faster.
x86 is not CISC. x86 is x86, it has its own weird style. For a good example of CISC design, examine the 68000 family. I’m still wondering whether the 68000 could have reached the same performance levels as the x86s if a real competition had continued. The RISC avantage of the early ’90s which killed the 68000 ( PowerPCs and ARM competing with x86s and 68000 ) disappeared when millions of extra transistors were possible. When you compare PPCs and x86 instruction sets, the PPC is superior in every aspect except code density ( which relates to cache size and memory bandwidth ). 68000 has a quite dense code.
Maybe IBM has suppressed the MMU ( paged/virtual memory support ) in the game console CPUs to get extra speed in the address path. It would predate the re-use in a real computer but can make sense in a closed design.
You just have to define the right metric (say, average increase in life expectancy in controlled groups consuming each product).
So you can compare apples to oranges after all,when you have the right referential fit in terms of metrics and so on?I doubt wether you get anything meaningfull out of the equation.
Well, Apple and IBM seem to think that you can compare performance, as they’re both members of the SPEC consortium.
How does Apples gaming performance fit in this context
(Doom3 etc) or is it now the GHz myth?
No, not at all. I’m not saying there is anything wrong with using a particluar system, but simply doing it to be different is just moronic.
Use it because it works better, because you prefer it over anything else, but if you have such a strong need to be unique that you run a system just because not a lot of other people do, then you have issues.
If Apple would switch to Intel, they would switch to Itanium and not x86s.
That would make no sense, because then they’d still be stuck with a single low-volume supplier who’s got trouble keeping up with x86. The fact that Intel itself is of course the Itanium’s main competitor only makes things worse. Also, due to the Itanium’s even more severe code bloat the technical merit of such a switch is debatable.
Competing with PCs doesn’t make sense.
True. But Apple wouldn’t be building standard PCs, they’d be building Macs with x86 processors.
I’m still wondering whether the 68000 could have reached the same performance levels as the x86s if a real competition had continued.
I’m sure it would, it’s sad Apple and Motorola fell for the RISC hype back then
.
The 68000 was a truly great architecture. Compared with x86, it had a much more coherent design. In fact it had many of the advantages of RISC architectures, e.g. a large(ish) register file and an orthogonal instruction set that make things easy for compilers, while avoiding RISC problems like the instruction-wasting load-store idea or the code bloat due to a fixed 32-bit instruction format.
The 68060 showed were the journey could have gone. Like the original Pentium, it was able to execute two instructions per cycle and also achieved similar clock speeds. The next step would have been an out-of-order implementation similar to the PentiumPro, which still serves as the basis for today’s Pentium M.
The 68000’s instruction set was well prepared for a 64-bit extension too.
Most instructions had a two-bit size field indicating the operand size: ’00’ meant 8-bit, ’01’ meant 16-bit, and ’10’ meant 32-bit.
’11’ was left reserved, with the obvious intention of someday using it for indicating 64-bit operands. No AMD64-style legacy/long modes or prefix acrobatics required.
And all that in a design from 1979.
finally jobs understands that is time to change his business stragedy, because apple with a slick OS, tower ,monitor is not doing as good where it boils down to the profits. when a CEO or president sees that a company for instance like Dell has a higer revenue then his such by using a 4 year OS called windows xp and OEM hardware that is found on the store shelvs for a low price and a cheap plain jane pc tower is making way more cash. that with a doubt shows that apple is and always have been doing bad business tactics. Sure tiger is nice but who care it’s not bringing in dell size of revenue to apple. YO jobs apple can make microsoft’s kind of revenue if you let people build their own pc and use the mac os.
I can picture an AMD Opteron in a mod case of my choice running the MAC OS playing couter strike online without worries of viruses because of the BSD kenel and the UNIX design. I think by letting people choose their own hardware and install mac os that alone can bring in a ton of revenue. I would like it if apple would switch to a gaming and desktop company for instance like having live internet play held by apple and other
interactive stuff. that is the direction that microsoft is heading but i would to see apple get there first
@rayiner
I don’t care how eloquent you try to prove your case about comparing anything, you just cannot. SPEC, that is like comparing Apples an Oranges on the basis that they both provide vitamin C, and then measuring the amount of vitamin C and claiming one is healtheir than the other. Thus, ignoring the benefits of calcium, and other vitamins and minerals that could state a case for the fruit that didn’t have as much vitamin C. Too many differences to be comparable.
@nimble
just because an argument sounds good, or seems wel reasoned, doesn’t make it correct. His argument was based on a lot of conjecture, and I just pointed that out.
Who cares is Apple is a member of SPEC, that still doesn’t mean squat. Just because the USA, and China are members or the UN, doesn’t mean that they share the same goverment, religion, morals, GDP, etc. Again, countries are tricky to compare, and for the most part, you cannot accuratly compare them. Too many differences again, and while you can try to compare them, most of the results will be based on assumptions.
@conny
you should be ashamed of yourself. Isn’t it the right of a person to choose to use whatever they want, for whatever reasons they want? Do you use linux? Guess all those people who worked on linux, and used linux in it’s infancy, because it wasn’t an MS product, were pathetic? It comes to choice, and some people want to be unique, and use hardware/software because it is different. We all just don’t wake up in the morning, and jump on the next bandwagon.
Remember, the most fundemental freedom, the one that all other freedoms are derived from, the only one granted to us by god, is the freedom of choice. Who are you to tell someone that their choice on buying a computer, because it is different, is wrong.
Now, if you were telling someone that their choice of killing someone is wrong, then you do have a good point there.
“finally jobs understands that is time to change his business stragedy, because apple with a slick OS, tower ,monitor is not doing as good where it boils down to the profits. when a CEO or president sees that a company for instance like Dell has a higer revenue then his such by using a 4 year OS called windows xp and OEM hardware that is found on the store shelvs for a low price and a cheap plain jane pc tower is making way more cash. that with a doubt shows that apple is and always have been doing bad business tactics. Sure tiger is nice but who care it’s not bringing in dell size of revenue to apple. YO jobs apple can make microsoft’s kind of revenue if you let people build their own pc and use the mac os.”
A higher revenue doesn’t mean that you aren’t making profits. Apple is actually is very profitable company, and while Dell has a higher revenue, that doesn’t mean they have a higher profit (but I would assume that Dell does).
just because an argument sounds good, or seems wel reasoned, doesn’t make it correct. His argument was based on a lot of conjecture, and I just pointed that out.
Yeah, great. But you’re not providing any arguments that actually disprove Rayiner’s assumptions or conclusions.
Who cares is Apple is a member of SPEC, that still doesn’t mean squat. Just because the USA, and China are members or the UN, doesn’t mean that they share the same goverment, religion, morals, GDP, etc.
So you don’t like comparisons, but you have a penchant for dodgy analogies anyway.
The SPEC website says: “The goal of SPEC is to ensure that the marketplace has a fair and useful set of metrics to differentiate candidate systems.” Why would Apple be a member of SPEC, if they didn’t agree with that goal?
The UN, by the way, isn’t about nations’ internal affairs, it’s about trying to maintain peace and get nations to cooperate.
What do you people classify as a mac? PPC architecture does not make it a mac. A mac is PPC + mac software, even more so a mac can be classified as just mac software regardless of hardware. But never can you classify the XBOX360 as a mac. The first XBOX was said to be a customized win2000 OS and you can tell that the XBOX360 is an even more customized windows OS since MS has ported their OS (NT) to PPC in the past.
A image conscience company such as MS would never use BSD, Linux, MACOS, or any other OS to promote their products. That kind of thinking is just asinine. Their media center initiative to wire your home should tell you that everything they do revolves around the windows OS.
“OK, it’s your opinion, but please don’t cry if the next Windows will kill Mac OS X.
Anyway… It’s ridiculous. Are you sure that the biggest software company can’t come up with something really good? Is it good to be a rabid fanboy?”
I really don’t care, although I think the idea of any single operating system “killing” another is pretty laughable. People who use Mac OS X to begin with do so because Macs fulfill their computing needs where PCs don’t. (And vice versa) It doesn’t matter how good Longhorn is, Mac OSX, Linux, BSD, and other operating systems will continue to exist. Conversely, it doesn’t matter how good any other operating system is, Windows will always exist for those who prefer that operating system. No need to cry about it, its called free market economy.
What will “kill” Mac OSX (and any other operating system for that matter) is when the company behind it loses touch with its user base. This was partly responsible for putting companies like Be Inc under.
As for being a “rabid fanboy” why don’t you get over yourself. It sounds as if you are suffering from the classical Freudian tendency to project your personal shortcomings onto other people. Get help.
You want to use Windows, be my guest. Use whatever you want and I will do the same. Just don’t expect Longhorn to be some magic bullet that is going to kill off ever other operating system. Not going to happen.
And I will freely admit that Microsoft could conceivably come up with something useful for the computer. After all, we all know how warmly their web browser-integrated-with-the desktop in Win98 was received. :-p
I think if Apple is going to port over to another processor architecure it would be Cell and NOT X86. What does Intel have to offer over Cell?
Low Price? NO
Performance? NO
The ability to run PowerPC binaries? NO
Low heat? NO
Which is more probable? Apple talking to Intel about WiMax or moving processor arch to X86?
I think Apple should port OSX to that new Cell Processor
Spending tens of millions of dollars to implement in a deliberately incompatible way the same functions ( AGP & PCI, HD controller, DRAM controller … ) and for the same performance levels as the cheap ( hi volume ) PC bridges available from Intel, AMD, VIA, … would be a _very_ bold move.
The graphics engines, LAN, USB, FireWire controllers are already the same between MACs and PCs. The transition occurred many years ago when Apple switched in the early PowerMacs from NuBus to PCI.
( Sun did the same from S-bus to PCI. The only important difference between a PC, A PowerMac and a Sun Ultra workstation is precisely the CPU ! )
>you should be ashamed of yourself. Isn’t it the right of a >person to choose to use whatever they want, for whatever >reasons they want?
Yes of course its the right of anyone to use whatever system for whatever system they want, however simply choosing because they want to be unique has to be the worst reason ever.
>Do you use linux? Guess all those people who worked on >linux, and used linux in it’s infancy, because it wasn’t an >MS product, were pathetic?
Not at all, because its not the same thing. Linus and the rest of the gang didn’t write/use Linux simply because they felt DOS/whatever was to mainstream. They did it for other reasons.
>It comes to choice, and some people want to be unique, and >use hardware/software because it is different. We all just >don’t wake up in the morning, and jump on the next >bandwagon.
I agree, its about choice, and since you have the right to choose a system for the sole reason of feeling special, I have the right to find that reason incredibly stupid. I am NOT saying that using another system is stupid. You feel more comfortable in windows? Use that. You prefer the UI of OSX? Use that instead, they’re both sound reasons.
>Remember, the most fundemental freedom, the one that all >other freedoms are derived from, the only one granted to us >by god, is the freedom of choice. Who are you to tell >someone that their choice on buying a computer, because it >is different, is wrong.
Well, as stated earlier, it’s my right. Who are you to tell me that my opinion in that matter is wrong then?
Seems to me that you think I’m claiming some system is better/worse than another system, I can assure you I’m not.
They are in two of the three next generation game consoles – which will have far more shipments than Apple machines. If I were Intel I would be trying to get Apple, since they lost XBox.
never said you didn’t have the right to voice your opinion, cause you do. All i am saying, is that you shouldn’t be telling people they are wrong for wanting to be unique. That’s all. You have that right, and they have the right to decide why they want to buy something. It might seem stupid to you, but might mean something to them.
In my opinion, I don’t mind people buying stuff to be unique, since, in the modern era of marketing, everyone buys the same stuff to be cool. Hell, look at the crap we call music today, that is forced down our throats, and, people feel they have to buy it (or download it), cause everyone else is.
apple talked with intel.
i talked with the nice postgirl this morning, too.
You posit that the comparison is in the CPU, not the product, which I believe a false assumption.
Apple does not cell CPUs, Apple sells products. So regardless of how they do it, if Apple labels it 2.7 GHz, and thats the clock speed that shows up in the System Profiler, then its a 2.7 GHz product.
Intel on the other hand, sells CPUs, and a lot of them, and companies like Alienware overclock them, and thats a valid comparison, because like Apple, Alienware sells products, not CPUs.
I guess it would amaze a lot of people to find out that Apple also talked with Asus in the past since Asus only makes PC motherboards(NOT)
Spending tens of millions of dollars to implement in a deliberately incompatible way the same functions ( AGP & PCI, HD controller, DRAM controller … ) and for the same performance levels as the cheap ( hi volume ) PC bridges available from Intel, AMD, VIA, … would be a _very_ bold move.
They wouldn’t have to do any of that, all they’d need to do is a PC-incompatible boot ROM, e.g. based on OpenFirmware which they have in Macs already.
I don’t think Apple would be overly concerned about people developing hacks to get Windows running on such Macs anyway. In fact, it might even be a selling point.
What they would be concerned about is other firms building OSX-compatible PCs. But just as Apple do at the moment, they could simply refuse to license OSX for anything but Apple hardware. As a consequence there could be no pre-installed non-Apple Macs and only geeks would find ways around that.
I remember, a long time ago, there were 68k Mac emulators ( The same way as Wine is an emulator ) for Atari ST computers ( Spectre and Aladdin ).
I can imagine someone re-implementing the MAC ROMs without actually copying a byte of Apple code, like Wine is re-implementing Windows APIs, but on a very low level.
In such a move, it would be hard for Apple to prevent people installing x86 OSX on their pcs.
Yes, of course you can buy a OS X license without a Mac, it costs $129.
But one of the first clauses of the MacOS license says:
This License allows you to install and use one copy of the Apple Software on a single Apple-labeled computer at a time.
So you can’t legally use MacOS on non-Apple hardware, or even in an emulator like PearPC.
Some people of course will be happy to ignore that and work their way around any technical hindrances. And you can’t really blame them for it, if have shelled out the $129. Apple might even welcome getting geek mindshare in this way.
The important thing for Apple is that no-one can sell computers with MacOS preinstalled. Most people don’t want anything to do with “installing an operating system”, even more so if any hacks are needed.
I think Apple should pair up with PearPC and make a stripped down version of OSX that is tightly integrated with PearPC so Dev’s can try stuff out on Mac’s