Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Mago

macrumors 68030
Aug 16, 2011
2,789
912
Beyond the Thunderdome
There's only theoretical support for ARM as it stands now. Sore, everything is avery benevolent world is just a compile away and that's real easy.. but in the real world, it's not that simple at all. Look at the support for very mature platforms like PowerPC.. IBM and Freescale tried for years to get the x86-centric Open source community to just "recompile".. but hey, it wasn't that easy at all. And IBM does make servers. One of the problems they faces was that x86 applications does not just recompile to big endian plattforms, and hey.. ARM is big endian. And I'm sure that Apple developers are eager to support another plattform, which would be very low volume and quite high in support. The toolchain is there, but would they jump eagerly to ARM just because Apple tries to release a niche server plattform? Hardly.

You're outdated, actually it's far mote complicated yo jump from x32 to x64, Xcode programmers do that w/o cry w/o hassle, switch from endian affects few bitwise related operations and the compiler it's self handles it automatically most cases, very few applications requiring low level bitwise operation (as cryptography) years ago coded the provisions to switch from endians (actually an objectiveC code on bitwise from powerPC can be compiled targeting ARM w/o editing nothing more than the header).

Major challenge from migration on OSX server from Mac x86 to ARM headless Server is to deal with Web based user interface than a Windowed one (notwithstanding an Apple appliance based on a7 could handle an rich graphical interface).

You know nothing about what I know, thank you.
We can see how software vendors are jumping to the opportunity to support AMD's, TI's, APM's and Cavium's ARM plattforms. No, they are not.. some products are there in theory, but hey, most are just prototypes, because its easy to just recompile for prototypes.. So

LAMP, mysql , postgress support to ARM was complete years ago LAMP applications run transparent either on x86 or arm, seems you need to do a tour at Hostgator or Godaddy

.. Oracle Java? Seems like Java v8 for 64-bit ARM is at least half a year away. Since that was a long time coming, this transition cannot really be considered a simple recompile. And why on earth would Oracle support Apple's tiny tiny platform?

Oracle Java by itself it's dead, and not as easy to migrate something coded on property platform or IDE if you want to support another platform, Oracle markets are vertical, they don't have nothing to look at LAMP market, mysql / postgress programmers will never come back to Oracle, so no sense to Oracle to launch an product on a platform aimed at users eager to use they solutions.

Ohh, and can you get MySQL support on ARM, or even a download? Or should a just recompile it myself?

Mysql migration to ARM platform it's old, actually older than the iPhone itself (you can run mysql databases on an iDevice), ignoring that tells me al lot from who are you.

Dell and HP are really throwing themselves behind their ARM endeavors.. Those Moonshot servers are really flying of the shelves. They must be since the most important thing is power savings. Or.. it might be x86 support. I don't know.. My money is on the latter, in the 1-10 year timeframe. And that's why the Moonshot servers are x86 now, and not ARM.

Actually you can do with your money what you deserves (sorry if sound loud) , competitive business can't do that, especially where the competition is so intense as the server farms, in fact Google is using ARM to run they platform (Google that sir).


And if recompile is such an easy thing to do.. why is the software support for mature server plattforms like POWER and Sparc nowhere near x86? And if we are talking performance/watt, single core performance and scale out parallelism.. neither ARM not x86 is near these guys. But still.. they are not particularly successful in the server space.. because of x86 legacy.

PowerPC is not an mature platform, it's what we call a legacy platform, same is for Sparc, don't need to comment more on that, only that one thing is to migrate an solution from platform if you need to migrate your IDE, not the case of MAC applications since Apple does the work to migrate Xcode (the ide) to target platforms you don't need even to care about (maybe only if you do a bit of magic manipulating directly data thru bitwise operations).

I'd certainly like to see that, but Apple really doesn't need to. Should Apple go don't this route, they could just pick an off the shelf ARM SoC, and run iOS on it. this is not the server we are talking about here, this is a slightly more capable Time Capsule, that's a far cry from Mac OS X Server..

No, I'm talking an appliance that actually offers all the services delivered by std osx server, it's easer to migrate osx server and recompile services and applications than migrate such applications and services to iOS.

I thought we were talking about servers.. not NAS. If were talking NAS, I'm with you.. go ahead. I want Apple to open up an iSO based SDK and App Store for a slightly modified Time Capsule and AppleTV. I really don't know why they haven't.

Actually those NAS are servers, people here developing and collaborating use them to host timemachine, mysql, lamp, apple shares, etc it's an growing market that's sucking both timecapsule and base osx server.

----------

Apple got out of the XServe business a long time ago. Why would they go back, and if they did, why in the bloody heck would they choose ARM?

You talk about a sort of mini Mac mini, something less powerful for those who need a truly low end server, but Apple is not in the business of selling cheap super-low-end hardware. They can, and do, convince people to buy a full Mac mini all the time. The margins are far more better on a Mac mini than what you describe. Apple can just plug in Intel-made chips on a small underpowered board and go to town. That's a lot easier and fits their business model better than designing some custom ARM processor, porting over all their software to ARM, and then trying to sell it for pennies.

When will you ARM cultists give up?
What Apple? The one what will never sell an rounded Mac Pro, or an oversized iPhone?
 
Last edited:

repoman27

macrumors 6502
May 13, 2011
485
167
You're right about everything except Thunderbolt. Thunderbolt is now DEAD. This just killed it. It's a total replacement for Thunderbolt 2 while retaining USB connectivity and cable standards AND is reversible. It gets rid of ALL major problems and ALL major complaints and it's a HUGE standard (unlike Thunderbolt which hardly anyone supports or uses except Apple). It sucks to be people who bought into Thunderbolt and it sucks to own a Mac Pro. You've just been PWNED! :eek: :D

I called this a couple of years ago. No one gives a crap about APPLE standards and Intel or not, it's ONLY been really used by Apple. You CANNOT create "standards" when you represent 8% of the market. Making matters worse, Apple wanted to have exclusive access to Thunderbolt the first year. WHY???? There was NOTHING available for it and it only meant that PC adoption would not occur for at least another year to even start. They helped doom their own connection standard. Call it Steve's dying folly if you like, but it was a stupid move.

Dream on. They can't even hardly get any devices to support Thunderbolt I several years later and they're all universally OVERPRICED. Don't give me the "Pro" line of defense because Apple has done everything in its power to KILL its own Pro market and hand it lock, stock and barrel to Windows. The Mac Pro is really a Prosumer machine (given the lack of internal expansion, rack mount capability and even software now that most have left the market), except most prosumers can't afford it. I doubt you'll even see Thunderbolt III. For goodness sake, look how long it's taken to get significant USB3 adoption by consumers (given the lack of desktop sales and the lack of mobile support). By the time Thunderbolt III makes an appearance, USB 4 will be out in short order and that will be the last you will EVER see of Thunderbolt, virtually guaranteed and IMO good riddance. We don't need competing standards and we certainly don't need overpriced products for those standards. Sometimes competition is BAD (look at HDDvd vs. Blu-Ray and how that only slowed adoption).

You never fail to amuse, Magnus.

You do realize that the Type-C connector is also likely to be the new Thunderbolt connector, and that the Alpine Ridge Thunderbolt 3 controllers will introduce a USB 3.0 signaling mode? They will also be available at least a year prior to Intel integrating USB 3.1 into their chipsets.

It does not really suck to own, for instance, a Mac Pro and a LaCie Little Big Disk Thunderbolt 2, which have been able to provide 1385 MB/s sequential reads from a single DAS unit since January of this year. Waiting until late 2016 for Intel to integrate USB 3.1 into their Union Point chipsets alongside Cannonlake might suck for most people. But the upside is that everyone will finally have a bus that might be able to crank out the same single DAS device speeds as OG Thunderbolt has since the beginning of 2011.

I think Apple is considerably more interested in creating points of differentiation between the capabilities of their hardware platforms and those of their competitors, rather than in simply making new standards.

Don't get me wrong, I'm very excited about the prospects of the USB 3.1, Type-C Connector and Cable, Power Delivery 2.0, and Billboard Device Class standards. You seem strangely upset about Thunderbolt not meeting your particular needs or budget, though. Well, have fun raging out until SuperSpeedPlus devices actually make it to the market in volume.
 
Last edited:

Mago

macrumors 68030
Aug 16, 2011
2,789
912
Beyond the Thunderdome
I think Apple is considerably more interested in creating points of differentiation between the capabilities of their hardware platforms and those of their competitors, rather than in simply making new standards.
I agree with you, this USB 3.1+TYPE C+DP3 delivers a lot of capabilities without the disadvantages of an single platform interface (Thunderbolt - PC Architecture), usb3.1 and dp3 can be deployed either on phones, tablets, appliances or pcpc, providing an common place for such interface Apple doesn't need to control (as the iPhone host connector), I don't agree Apple regrets open std, few interfaces are Apple exclusive and I consider Apple has legit interest to control it.
 

Mago

macrumors 68030
Aug 16, 2011
2,789
912
Beyond the Thunderdome
In general, you're right though. Thunderbolt is a honking big data pump that provides much higher serial I/O bandwidth than any of it's contemporaries, combined with protocol flexibility.

I think you miss understood the purpose of Thunderbolt and why it's different and will live long on Intel platform.

Intel purpose in Thunderbolt never was to monopolize peripheral interface; but to provide an std expansion interface communicating devices without need to write new drivers as long those devices previously existed on an pcie version, that's why we see a lot of storage solutions based on Thunderbolt they just *plugged* an sata - pcie host chip to one end of the Thunderbolt chain, from the peripheral pov it's the same as being connected to the logic board on some expansion board, of course high performance it's another capability plus unparalleled bandwidth efficiency since Thunderbolt lacks protocol overhead as usb.

Thunderbolt big advantages are high bandwidth, common device drivers for internal or external peripherals, the disadvantages are of course: cost, besides the Thunderbolt interface you need an pcie capable peripheral that's why TB devices use to cost near 100$ more than usb3; another big disadvantage is you can only daisy chain 7 devices while on usb you can connect upto 255 device to a single head.

In no way the USB3.1 kills Thunderbolt, of course those applications demanding bandwidth and support for an legacy pcie device will go one Thunderbolt since saves a lot of money on R&D and time to market (as the case of Thunderbolt->m.2 storage) , whatever I don't see future to Thunderbolt on personal portable HDD (spinner) since no spinner HDD has bandwidth enough to require 1/5 of Thunderbolt bandwidth, usb3 actually provided superior convenience and imperceptible/unexistent bandwidth bottleneck.

Thunderbolt will be soon a niche interface for extreme bandwidth hungry devices as external m.2 ssd, and other interface not available on USB.
 

deconstruct60

macrumors G5
Mar 10, 2009
12,296
3,890
....
However, Apple absolutely did have an exclusive on Thunderbolt for the entirety of the Sandy Bridge era. ... An Anandtech article from shortly after the embargo lifted can be found here. This was not a rumor, and was never denied by Apple or Intel.

Why would the deny something that wasn't true? Exclusive has a meaning when it comes to commodity/things/news stories/etc. that connotes only available from one source. Not that it is technically contractually bound to being sold to just one vendor. That articles sentence is likely being pulled out of context to imply that Apple/Intel actively blocked other vendors from getting TB. Apple being the only vendor who completed the certification would make them sole vendor, but it isn't because the others are being blocked.


Several flaws with Apple locking out everyone else contractually (i.e., Intel would only sell controllers to Apple). .

First is the opening sentence of that article:

"...Intel has informed us that we will see updated drivers for Windows certified Thunderbolt devices that will enable hot plugging under Windows as well as address some of the behavior we've seen. ..."

Thunderbolt without hot-plugging isn't Thunderbolt. It is never going to pass the certification tests. If can't pass the TB certification test , what high incentive do Windows PC system vendors have for shipping a system? .... Slim to none.

I think Apple did play a role in influencing that the certification process for Thunderbolt would be quite rigorous. Apple has little interest in seeing a race-to-the bottom occur on TB. Wonky boxes attached to Apple systems are bound to raise Apple's support costs. Getting into a race-to-the-bottom with USB is a dead end game that neither Apple or Intel wanted to see TB get into.

From the software perspective the exclusivity that Apple has here is what they always have with Mac OS X system. Only Apple hardware can be legally paired up with OS X.

But, but Sony's hack got released....

That is the second reason. Sony's USB plug system along with the Light Peak demo with what appeared to be USB plug is very indicative that somewhere along the way the design changed. So this is a new tech that is flip-flopping on plug design and doesn't have OS support to pass certification. Yeah sure, Windows PC vendors are going to be chopping at the bit to dump $100k-1000K into systems R&D. Probably not.

AFter Sony got burnt with the non standard Thunderbolt device what did they do with Thunderbolt after that? Not much. Sony getting burt only reinforces the risk averse stance the PC system vendors would take.


This software/hardware interface issue is still in play. Third factor in the reasoning.

"... These have to hang off the platform controller hub (PCH). On the other hand, Apple was allowed to hook up the Thunderbolt silicon directly to the CPU. The reason behind this leads us to the software side of things. ..."
http://www.anandtech.com/show/8529/idf-2014-where-is-thunderbolt-headed

Currently, Apple has an exclusive in systems that have the TB controller(s) hanging off the CPU. Contractual blocking? Nope. Once again Apple's OS X is out in front of Windows support. If the folks in Redmond were aggressively pursing TB as much as the folks in Cupertino, Apple would have any exclusivity. As long as they do, then Apple will get to market before the systems dependent upon Windows.

The notion that Thunderbolt was completely software transparent ( the old software will just work) was only just a bit smaller in marketing hype than the position that TB was the one port to rule them all. It never was. There isn't a huge chasm or vastly new API to cross. PCIe has optional hot-plug features. The issue for TB is just that most folks don't implement them either in the OS or in specific equipment drivers in the mainstream PC market.

Circa rush of discrete USB 3.0 controllers that need extensive xHCI support versus assigning resources to Thunderbolt it is relatively not surprising that most Windows PC system vendors went in the direction where Microsoft was sinking the most of there emphasis on ( USB 3.0 / xHCI). Defacto that handed Apple exclusivity for a period of time.
 

randian

macrumors 6502a
Jan 15, 2014
784
362
Sony's USB plug system along with the Light Peak demo with what appeared to be USB plug is very indicative that somewhere along the way the design changed.
Not so sure about that. Sony has a history of trying to make a standard into something proprietary. Look what they did with Firewire, for example. A non-standard, Sony-only version.
 

deconstruct60

macrumors G5
Mar 10, 2009
12,296
3,890
Apple switched to Thunderbolt first because Intel was slow to adopt native USB 3 on their CPUs and Apple didn't want to add support themselves to the CPUs.

No.

1. Intel hasn't added USB 3 support to the CPUs (putting aside the SoC models for a moment). USB 3.0 support is added to the PCH chipset. The chipset is bundled for sale with the CPU ( it both are useless without the other) but it is typically two chips.

There are now some "CPU" that Intel sells that are closer to being SoC units where the PCH die is placed in the same "CPU" package as the x86 core (+ non-core elements) die.

2. For many Macs Apple could have gone with a discrete USB 3.0 controller. The current Mac Pro has one and the sky hasn't fallen. It is cheaper (both in chipset costs and R&D costs ) for Apple to pick one and only one USB 3.0 controller and just it everyone. It was not necessary.

3. Thunderbolt isn't a USB replacement. They overlap in capabilities a bit but one is not a replacement for the other. Far more pressing for Apple was their quest to get rid of ExpressCard on laptops and to deny any very high speed I/O expansion capability at all to the iMac and Mini. Thunderbolt helps solve that issue. It also moves the market penetration forward on the mini-DisplayPort standard that Apple created.

USB 3.0 devices would just work slower with Apple's USB 2.0 limited ones but that was just a temporary bump. Between FireWire 800 and Thunderbolt many of the systems Apple sold in that interim year had alternatives to being suck on USB 2.0. Apple could have also swapped out Firewire controller for discrete USB 3.0 controller. That probably would cause more drama than it would be worth.

4. Apple was also probably waiting in part on native USB 3.0 support in Windows. Both OS X and Windows native xHCI support didn't come "sooner" rather than later.



Before this announcement I would've added if you want a simple, convenient dock for a laptop, go Thunderbolt.

Eventually, Type-C will probably allow for cheaper (than TB ) docks for those who don't need much more than ~50-70W power and a display connection. I suspect future TB controllers may be able to deal with this by skipping TB and just pushing USB 3.0

http://arstechnica.com/gadgets/2014...olt-doubles-speeds-but-changes-the-connector/

Again Thunderbolt 3 isn't technically coupled to Skylake ( the CPU) but with the infrastructure (i.e., chipsets ) coming with Skylake.

The "smaller" connector that does USB , Display Port, and HDMI with adapters sure sounds alot like a derivative Type-C connector with a different controller on the other side.
 

MagnusVonMagnum

macrumors 603
Jun 18, 2007
5,193
1,442
You never fail to amuse, Magnus.

I'm glad I'm hear to entertain. :D

Don't get me wrong, I'm very excited about the prospects of the USB 3.1, Type-C Connector and Cable, Power Delivery 2.0, and Billboard Device Class standards. You seem strangely upset about Thunderbolt not meeting your particular needs or budget, though. Well, have fun raging out until SuperSpeedPlus devices actually make it to the market in volume.

I'm not upset. I'm happy to see Thunderbolt go in the long run. I never thought it would work out from Day 1 and so far I see nothing to indicate I was wrong. It's overpriced in terms of current offerings versus USB3 (e.g. hard drives of the same size and speed where Thunderbolt offers no advantage), has poor adoption, disagreements over the end connector (Sony choosing a USB style connector, although I don't know if they ever even bothered to offer it thus far) and has little chance against USB for obvious reasons (market adoption in mass numbers). Given USB is obviously an able competitor if only previously behind a bit on the timetable, I simply see zero need for Thunderbolt. If it were 10x faster or 5x faster, it would impress. But USB 3.1 is exactly the same speed as Thunderbolt II (not one, but two). It's caught up. Thunderbolt is moot. going 2x faster isn't fast enough. USB 4 will be that fast or faster. If there was a huge need, it would get to market faster as well. But few need it.
 

deconstruct60

macrumors G5
Mar 10, 2009
12,296
3,890
Not so sure about that. Sony has a history of trying to make a standard into something proprietary. Look what they did with Firewire, for example. A non-standard, Sony-only version.

Not really. First, Sony's variant was adopted into the standard. Sort of like how mini DisplayPort was weaved into DisplayPort standard later. Second it was a different format. In Sony's usage the picked an existing physical standard and varied the protocol on the wires a bit ( similar to what Light Peak did). Sony did what Intel had already done (and in part leveraged the "fiber" USB 3.0 work done by Intel and others previous that was rejected by the USB-IF. At one point USB 3.0 was going to have options that covered some of the things Light Peak was suppose to do. ).

There are times when Sony has been way out there detached from everyone else. This particular case far more looks like trying to lever sunk costs had put into efforts that the "standards" veered away from.
 

deconstruct60

macrumors G5
Mar 10, 2009
12,296
3,890
Once the networking stack gets optimized (all done in software right now so IO can affect performance) then thunderbolt can take off as a networking technology too (also need switches). Remember too thunderbolt is specced to go to 100Gbps and USB 3.1 is just catching up to Thunderbolt 1

Thunderbolt is a switch so it doesn't particular require more switches layered on top. Thunderbolt is a simple network topology which is why the internal switches are cost effective. There is little to no likely advantage to looping in external switches into the networking mix. An completely arbitrary network likely would make the external switch cost as much as the 10GbE and the like you are trying to avoid.

Thunderbolt network is only likely going to be useful as a point-to-point solution that works for less than 4 sized workgroups.

Thunderbolt is primarily a direct attached network.
 

repoman27

macrumors 6502
May 13, 2011
485
167
Why would the deny something that wasn't true? Exclusive has a meaning when it comes to commodity/things/news stories/etc. that connotes only available from one source. Not that it is technically contractually bound to being sold to just one vendor. That articles sentence is likely being pulled out of context to imply that Apple/Intel actively blocked other vendors from getting TB. Apple being the only vendor who completed the certification would make them sole vendor, but it isn't because the others are being blocked...

Where is this even coming from? Pretty sure you were around in 2011... Apple collaborated with Intel to produce Thunderbolt, created the Thunderbolt active copper cable (and the miniDP connector before that), came up with the ridiculous marketing name and performed the trademark filings for it, and inked a deal with Intel to be the only OEM to produce PCs with Thunderbolt host controllers during the Sandy Bridge era.

I'll have to go digging through the piles of slide deck pdfs I have from Intel briefings on Thunderbolt to get it direct from the horse's mouth, but in the meantime I'll quote this not-so-ambiguous language from another Anandtech article:

Intel engineered the spec, but Apple helped with a lot of the connector and cable design and as a result received a year long exclusive on Thunderbolt...

With the exclusivity agreement over, Intel's partners in the Windows PC space are allowed to ship Thunderbolt enabled motherboards and systems.

2. For many Macs Apple could have gone with a discrete USB 3.0 controller. The current Mac Pro has one and the sky hasn't fallen. It is cheaper (both in chipset costs and R&D costs ) for Apple to pick one and only one USB 3.0 controller and just it everyone. It was not necessary...

I've thought about this a *lot*. While Apple has the billions to move mountains and thus could have equipped the 2011 Macs with discrete USB 3.0 controllers, the logistics of doing so would have been such that it was utterly impractical in real terms.

Apple shipped roughly 21.8M Macs during that period. In February 2011 when the first of the SNB Macs became available, there were only 2 companies producing certified USB 3.0 host controllers: Renesas and Fresco Logic. Fresco Logic (who Apple ultimately chose as their supplier) had only shipped about 1M controllers in total at that time, and Renesas had shipped around 20M over the previous 15 months but had only ramped production to 3M units per month. It was dubious at best that Apple could have secured the volume of controllers they required, especially considering neither of the certified products offered more than 2 ports. Renesas, by far the leader in production volume, would not have an xHCI 1.0 compliant 4-port design available until March, and Fresco Logic wouldn't get their 4-port FL1100EX out the door until September.

Oh, and then there's the part where Apple is also responsible for developing and validating their own USB 3.0 drivers for OS X. Seeing as the Intel xHCI controller was obviously the baseline target for compatibility, going off and developing a driver to deal with early silicon from Fresco or Renesas would have been a massive waste of resources at that point in time. The whole endeavor would have been even more fraught because Apple would have no idea during the early stages of development whether they could count on a single controller source or how long it might take the chip they bet on to pass certification.

Combine all this with Apple's tiny product stack, all or nothing attitude, and insistence on not shipping crap just to tick a box on a spec list, and you get no USB 3.0 for 2011 Macs. Those early discrete USB 3.0 controllers were also not terribly good from either a power or performance perspective, and simply would not have provided the kind of user experience that Apple strives for anyway.

I'm glad I'm hear to entertain. :D

I'm not upset. I'm happy to see Thunderbolt go in the long run. I never thought it would work out from Day 1 and so far I see nothing to indicate I was wrong. It's overpriced in terms of current offerings versus USB3 (e.g. hard drives of the same size and speed where Thunderbolt offers no advantage), has poor adoption, disagreements over the end connector (Sony choosing a USB style connector, although I don't know if they ever even bothered to offer it thus far) and has little chance against USB for obvious reasons (market adoption in mass numbers). Given USB is obviously an able competitor if only previously behind a bit on the timetable, I simply see zero need for Thunderbolt. If it were 10x faster or 5x faster, it would impress. But USB 3.1 is exactly the same speed as Thunderbolt II (not one, but two). It's caught up. Thunderbolt is moot. going 2x faster isn't fast enough. USB 4 will be that fast or faster. If there was a huge need, it would get to market faster as well. But few need it.

USB 3.0 SuperSpeed is 5 Gbit/s with 8b/10b encoding.
USB 3.1 SuperSpeedPlus is 10 Gbit/s with 128b/132b encoding.
Thunderbolt is 2x 10.3125 Gbit/s channels with 64b/66b encoding.
Thunderbolt 2 allows both channels to be bonded into a single 20 Gbit/s link.

A single Thunderbolt channel offers more bandwidth than USB 3.1 and 2.5x the bandwidth of USB 3.0.
A Thunderbolt 2 link packs more than twice the bandwidth of USB 3.1 and 5x the bandwidth of USB 3.0.

Until USB 3.1 is integrated into chipsets, it will come in the form of discrete controllers which are unlikely to offer more than a PCIe 3.0 x1 back end which will be connected to lanes coming from the PCH.
Thunderbolt 3, which will be available alongside Skylake, will be 4x faster than those USB 3.1 controllers on both the front and back ends.

Since 2011, Thunderbolt has enabled 1000 MB/s data transfers while simultaneously driving a 2560 x 1600 display from a single port, or 4 years before USB 3.1, the Type-C connector and the DisplayPort Alternate Mode will be able to do the same.

For as long as Apple needs Thunderbolt, it will be around.
 

Nightarchaon

macrumors 65816
Sep 1, 2010
1,393
30
Bye Bye thunderbolt, it wasnt at all nice knowing you :rolleyes:

Thunderbolt, expensive cabling and annoyingly priced equipment, or USB, cheap as chips...

i will bet now which one will be mainstream in a years time and which one will still be bloody expensive and niche product, if its not been abandoned by everyone but apple already
 

MagnusVonMagnum

macrumors 603
Jun 18, 2007
5,193
1,442
USB 3.0 SuperSpeed is 5 Gbit/s with 8b/10b encoding.
USB 3.1 SuperSpeedPlus is 10 Gbit/s with 128b/132b encoding.
Thunderbolt is 2x 10.3125 Gbit/s channels with 64b/66b encoding.

That type of difference will make no difference to the consumer at all as USB 2.x vs. Firewire 400 showed. Firewire 800 was an even bigger improvement for hard drive use and was the only game in town for those kind of transfer rates (and unlike current drives, most of which don't saturate USB 3.0 and so only cost more for Thunderbolt, it really cut transfer times by nearly half), most consumers STILL rejected Firewire 800. It cost too much and most probably didn't realize it made any difference since they didn't use external drives for high speed applications.

Thunderbolt 2 allows both channels to be bonded into a single 20 Gbit/s link.

A single Thunderbolt channel offers more bandwidth than USB 3.1 and 2.5x the bandwidth of USB 3.0.

None of which matters to the consumer. Short of the Thunderbolt 2 mode, which hardly anything uses, USB 3.1 counters everything else Thunderbolt offers from Display Port adapter combination ports, reversible cables (like Lightning), very similar speeds to the Thunderbolt found on most computers out there (save Mac Pro) and unlike Thunderbolt, it's completely backwards compatible with USB 3.x, 2.x and 1.x that are found on almost every computer made in the past 10+ years. A couple of esoteric video guys will use Thunderbolt 2. How hard would it be for USB 3.1 to take two chipsets and combine them into 20Gbit channel? There's no demand for it. Many people aren't even using USB 3.0 devices yet (your typical PC user just uses whatever the thing comes with and adds nothing and probably doesn't even do backups). The problem is you can't make a STANDARD based on 8% of the market, of which probably <1% is even using their Thunderbolt port at all, except possibly as a Display port adapter (not really Thunderbolt; it just shares the same port).

For as long as Apple needs Thunderbolt, it will be around.

I notice you say "Apple" rather than Mac users because hardly anyone uses it at all. It's a failure of a "standard" and will never even come close to Firewire market penetration, let alone replace something like USB. It's for high-end esoteric use only and how long will it even maintain that status? A few more years? Thunderbolt is a product in search of a market. If it didn't cost so darn much for the peripherals, things might be different. But who wants to pay 50% more for the same speed hard drive in Thunderbolt form compared to USB 3.x if the drive is the limiting factor? THAT is why it fails. Short of large RAID configurations of solid state drives for video editing, it's moot for most other devices as they don't need that kind of throughput. The fact you have to daisy-chain the devices is also a major annoyance, particularly in conjunction with monitors which often don't provide any kind of pass-through and so you have to disconnect your display while you fiddle with devices mid-chain. Awful.
 

toke lahti

macrumors 68040
Apr 23, 2007
3,270
502
Helsinki, Finland
First,
correct me if I'm wrong, but none of those articles has made tech specs clear. So is usb3.1+DAM 4 lanes with 2 lanes of usb3.1 (2x10Gbit/s) and 2 lanes of dp (2x8Gbit/s) and these will work at full speed with 1 meter passive cable or longer active cables?
You're missing the point that Thunderbolt isn't just a cable type, it's direct access to the motherboard.
Although TB theoretically is more elegant and aesthetic, than many "indirect" access, in the end, what counts is how many MB/s you get with one dollar AND how widely it is supported. Usb3.1+DAM (wonderful name for marketing ;) ) will win these both hands down within one year. Will be funny to watch how fast Apple will ditch TB. After all macs are not made for niche professionals.
Apple shipped roughly 21.8M Macs during that period. In February 2011 when the first of the SNB Macs became available, there were only 2 companies producing certified USB 3.0 host controllers: Renesas and Fresco Logic. Fresco Logic (who Apple ultimately chose as their supplier) had only shipped about 1M controllers in total at that time, and Renesas had shipped around 20M over the previous 15 months but had only ramped production to 3M units per month. It was dubious at best that Apple could have secured the volume of controllers they required, especially considering neither of the certified products offered more than 2 ports. Renesas, by far the leader in production volume, would not have an xHCI 1.0 compliant 4-port design available until March, and Fresco Logic wouldn't get their 4-port FL1100EX out the door until September.
All other major computer manufacturers had models with usb3 in 2011, Apple was the last in June of 2012. Renesas alone could have made all usb controllers for all macs, if Apple would have asked. You know, they can make more of chips that have high demand. Btw, AMD's A70M had usb3 integrated in June 2011.
If Apple, being the biggest muscle in industry had pushed usb3 harder after ratifying it in USB-IF in 2008, usb3 would have appeared on macs a lot earlier and also overall adoption of usb3 would have been a lot faster. Remember how much Apple affected in usb adoption in 90's, even when macs were even more rare than today? There were more than 100 million usb3 devices shipped before Apple shipped its first usb3 compliant product.
Interestingly, fluently working usb3 cards for macPro were available in the summer of 2011...
Instead Apple put their interest in TB, which I think wasn't very optimal choice. Since then Apple has been one gen behind in display technology and maybe they have had enough for now. It will be embarrassing for Apple to not support dp1.3 in their products in 2015 when all other new computer models will support it. Maybe Apple just puts hdmi2.0 in macs and acts like dp1.3 is not needed. Or some models will have TB and some models usb3.1+DAM. Can't believe that any other than nMP2 could have both (in 2015q3-4?). Keeping lightpeak and displayport separate would have been better choice for all these years. If only USB-IF (and Apple within it) would have accepted Sony's idea to combine lightpeak and usb. Now we are getting there with usb3.1+DAM many years later...
 
Last edited:

toke lahti

macrumors 68040
Apr 23, 2007
3,270
502
Helsinki, Finland
Btw,
if this is what future will be:
http://www.anandtech.com/comments/8558/displayport-alternate-mode-for-usb-typec-announced/423284
then this is finally the-one-connection-to-rule-'em-all.
It will be interesting to see how the prices evolve, those controllers will be expensive for starters just like for TB, but unlike TB, these might be in every device on the planet, so mass production can lower the price steeply and quickly. Not to mention that it will not be one company's monopoly...
 

repoman27

macrumors 6502
May 13, 2011
485
167
That type of difference will make no difference to the consumer at all as USB 2.x vs. Firewire 400 showed. Firewire 800 was an even bigger improvement for hard drive use and was the only game in town for those kind of transfer rates (and unlike current drives, most of which don't saturate USB 3.0 and so only cost more for Thunderbolt, it really cut transfer times by nearly half), most consumers STILL rejected Firewire 800. It cost too much and most probably didn't realize it made any difference since they didn't use external drives for high speed applications.

None of which matters to the consumer. Short of the Thunderbolt 2 mode, which hardly anything uses, USB 3.1 counters everything else Thunderbolt offers from Display Port adapter combination ports, reversible cables (like Lightning), very similar speeds to the Thunderbolt found on most computers out there (save Mac Pro) and unlike Thunderbolt, it's completely backwards compatible with USB 3.x, 2.x and 1.x that are found on almost every computer made in the past 10+ years. A couple of esoteric video guys will use Thunderbolt 2. How hard would it be for USB 3.1 to take two chipsets and combine them into 20Gbit channel? There's no demand for it. Many people aren't even using USB 3.0 devices yet (your typical PC user just uses whatever the thing comes with and adds nothing and probably doesn't even do backups). The problem is you can't make a STANDARD based on 8% of the market, of which probably <1% is even using their Thunderbolt port at all, except possibly as a Display port adapter (not really Thunderbolt; it just shares the same port).

I notice you say "Apple" rather than Mac users because hardly anyone uses it at all. It's a failure of a "standard" and will never even come close to Firewire market penetration, let alone replace something like USB. It's for high-end esoteric use only and how long will it even maintain that status? A few more years? Thunderbolt is a product in search of a market. If it didn't cost so darn much for the peripherals, things might be different. But who wants to pay 50% more for the same speed hard drive in Thunderbolt form compared to USB 3.x if the drive is the limiting factor? THAT is why it fails. Short of large RAID configurations of solid state drives for video editing, it's moot for most other devices as they don't need that kind of throughput. The fact you have to daisy-chain the devices is also a major annoyance, particularly in conjunction with monitors which often don't provide any kind of pass-through and so you have to disconnect your display while you fiddle with devices mid-chain. Awful.

Consider the number of Macs shipped by Apple in the past year. Consider the number of user accessible PCIe slots in those Macs. Apple has 100% committed to Thunderbolt as its sole PCIe expansion option for customers. At best, a USB Type C Alternate Mode could offer an x2 PCIe link. Unless you use a PCIe switch to borrow a couple of the PEG lanes, these will be PCIe 2.0 from the PCH until Skylake at the earliest. And the very earliest any Macs will ship with USB 3.1 will be late in 2016 when Intel integrates it into the chipset alongside Cannonlake.

If you want to proclaim Thunderbolt is dead, then you need to actually lay out a credible strategy and timeline for how Apple transitions from its current product stack to one that doesn't include 30M Thunderbolt devices annually. I said Apple because they are the only real customer for Thunderbolt, and they also happen to be one of the biggest semiconductor purchasers in the world. Thunderbolt will die when Apple no longer needs it.

Apple has managed to capture the majority of the profit from the PC, tablet and smartphone markets, with as little as 5% global marketshare in the case of Macs. They have also managed to amass one of the largest market caps in history for any company in any industry. Apple's success is in no way attributable to the lowest common denominator always triumphing over more expensive niche products.

If you're actually here with an open mind and not just trolling, then how does your narrative tie into reality in any way? What steps does Apple take to lessen their current dependency on Thunderbolt and when?

The other problem with your arguments is that you liberally commit the fallacy of accident. You can't just destroy the exceptions like that. There are people whose workflows include PCIe devices for which no USB equivalents exist. There are people whose workflows demand sequential performance from a DAS solution which is greater than USB can currently provide. There are people whose workflows require NAS / SAN throughput or connectivity that USB is currently unable to provide. There are people whose workflows require 10GbE or Infiniband networking capabilities. Some of these people are doing things with these technologies that are of great benefit to you, to me, and to everyone else. Thanks to Thunderbolt, they can do any of these things on a Mac if they want to.

I could, for instance, test a new USB 3.1 xHCI IP block on a PCIe x4 FPGA card in an external Thunderbolt enclosure. Certainly can't do that over USB 3.0. But darn it, you're right. The people building your future certainly shouldn't have the option of doing it on Apple hardware because Thunderbolt is too expensive for some people to afford and too complicated for some people to figure out. Maybe all PC's should be Chromebooks limited to accessing Google, Facebook and reddit? I guess we'll just have to put an opinion poll on the Internet to determine which arbitrary percentage of the population a technology must be suited for before it can be considered a success.

Thunderbolt may have woefully small attach rates with general consumers, but that in no way means it's a failure, because it was never really meant for them. For professionals, these devices are actually the tools they use to do their job. Your arguments are as asinine as saying that rotary hammers are dead because most consumers are just fine with a much less expensive and versatile hammer drill/driver and don't understand the differences anyway.
 

MagnusVonMagnum

macrumors 603
Jun 18, 2007
5,193
1,442
Consider the number of Macs shipped by Apple in the past year. Consider the number of user accessible PCIe slots in those Macs. Apple has 100% committed to Thunderbolt as its sole PCIe expansion option for customers.

You don't even realize that Thunderbolt is a step BACKWARDS if you want to look at it as an alternative to the previous Mac Pro's actual PCI expansion. Honestly, you seem to have missed that the Mac Pro market of the past has pretty much FLED Apple due to Apple not giving a flying crap about what true professionals WANT in a product. The current Mac Pro is NOT it and virtually NO ONE is using those ports in place of previous PCI products. Elvis has left the building.

If you want to proclaim Thunderbolt is dead, then you need to actually lay out a credible strategy and timeline for how Apple transitions from its current product stack to one that doesn't include 30M Thunderbolt devices annually.

You seem to confuse Thunderbolt capable computers with Thunderbolt devices and of those devices, how many would not work full speed under USB 3.1? Very few.

I said Apple because they are the only real customer for Thunderbolt, and they also happen to be one of the biggest semiconductor purchasers in the world. Thunderbolt will die when Apple no longer needs it.

Again, you state a FACT (that Apple is the ONLY one using Thunderbolt in any quantity and that they are the only ones that apparently "need" it since hardly anyone is buying Thunderbolt products that couldn't work under USB 3.0 or 3.1 and are 2-4x cheaper at that. In other words, CUSTOMERS don't "need" Thunderbolt. That is why you keep saying "Apple needs it" and not "Mac users" because WE do *NOT* need it. It's overpriced and not a substitute for an older Mac Pro with bus lanes that are much faster and more abundant.

Apple has managed to capture the majority of the profit from the PC, tablet and smartphone markets, with as little as 5% global marketshare in the case of Macs. They have also managed to amass one of the largest market

First you includes the entire mobile market in your sentence and THEN attribute it to only 5% Mac sales. WTF?!? Stick to one idea at a time. Apple controls less than 20% of the mobile market and less than 10% of the desktop market. PROFIT has NOTHING to do with adoption rates *OR* USAGE of a particular technology. Who cares how much money Apple is making in regards to Thunderbolt ADOPTION rates??? They have NOTHING and I mean NOTHING to do with one another. Your logic makes no sense AT ALL. Give it up. You can't make even a fair argument, let alone a good one.

If you're actually here with an open mind and not just trolling, then how

Ah, so you're losing the argument by leaps and bounds and you then drop the "Troll" card. This is where I stop talking to you because if you've lost the argument 100%. I am not a troll. I've been on here for almost a decade and own four Macs. How the hell am I troll? Because I don't agree with your opinions? That seems to be the most common argument these days for "trolling". They don't agree with me so they must be a troll. How can they not see things just like me? I hate everyone that's different from me. Yeah, that pretty much sums up ALL conflict in the world in its entirety save that caused by greed including all political, religious and taste arguments.

What steps does Apple take to lessen their current dependency on Thunderbolt and when?

What dependency? Apple already has USB 3.0 in its entire lineup now. They can add USB 3.1 quite easily. And if they really cared about PCI expansion, they should add an actual professional tower back into the mix instead of that freaking trash can look-alike that SHOULD have been the next Mac Mini or xMac instead of a replacement for a Mac Pro which it in NO WAY functionally replaces (and requires a load of outboard awful looking mess of equipment to even attempt to replace, assuming one doesn't need the higher speed PCI lanes. How is THAT better than a tower that lets you keep everything internal and neat? It's not. It's a freaking GIMMICK that really only works for video editors and very few others without a messy desktop.

So to answer your question, bring back REAL PCI expansion instead of a R2D2 look-alike where even the video cards use custom connectors so you can't even replace the video card without paying the Apple Tax.

Who asked for Apple to make a Mac Pro without PCI expansion? NO ONE!
 

repoman27

macrumors 6502
May 13, 2011
485
167
You don't even realize that Thunderbolt is a step BACKWARDS if you want to look at it as an alternative to the previous Mac Pro's actual PCI expansion. Honestly, you seem to have missed that the Mac Pro market of the past has pretty much FLED Apple due to Apple not giving a flying crap about what true professionals WANT in a product. The current Mac Pro is NOT it and virtually NO ONE is using those ports in place of previous PCI products. Elvis has left the building.

The Mac Pro (Mid 2012) only had two PCIe 2.0 x16 slots and 2 PCIe 2.0 x4 slots. If you ran a pair of dual-slot GPUs, you were only left with a single PCIe 2.0 x4 slot. The Mac Pro (Late 2013) has two PCIe 3.0 x16 slots which are permanently occupied by twin GPUs, and three PCIe 2.0 x4 slots made available via Thunderbolt 2.

The new Mac Pro limits you to just three AMD GPU options, which could be regarded as a serious step backwards if not an outright deal-breaker. It also offers WAY less flexibility than any other Ivy Bridge-EP platform I can possibly think of in terms of how the PCIe lanes are allocated. However, compared to any previous Mac Pro with dual-GPU's, the Thunderbolt configuration on the new Mac Pro actually provides more flexibility and more available PCIe bandwidth.

Apple rarely breaks out sales figures by individual product, but I'd be curious as to how the Late 2013 Mac Pro has performed sales-wise compared to any of Apple's previous workstation class Macs. Who knows, maybe you're right and the tube will go the way of the cube. I kinda feel like this has been one of the strongest selling Mac Pros in a long time though, but there are obviously quite a few factors playing into that which have nothing to do with the overall strength of the platform.

You seem to confuse Thunderbolt capable computers with Thunderbolt devices and of those devices, how many would not work full speed under USB 3.1? Very few.

Not at all. Apple sells Thunderbolt enabled PC's (Macs) as well as peripheral devices (ATD, Thunderbolt to GbE and FireWire adapters). See, what I did was estimate the combined annual sales of both classes of devices from publicly available data.

Not a single one of the millions of Thunderbolt devices sold since the interface debuted back in 2011 can possibly work at full speed under USB 3.1 until it progresses beyond the specification and FPGA prototype stage into products that people can actually buy. And PCIe devices without USB analogs will continue to exist and be the primary focus of Thunderbolt even after USB 3.1 finally achieves widespread adoption in 2017.

Again, you state a FACT (that Apple is the ONLY one using Thunderbolt in any quantity and that they are the only ones that apparently "need" it since hardly anyone is buying Thunderbolt products that couldn't work under USB 3.0 or 3.1 and are 2-4x cheaper at that. In other words, CUSTOMERS don't "need" Thunderbolt. That is why you keep saying "Apple needs it" and not "Mac users" because WE do *NOT* need it. It's overpriced and not a substitute for an older Mac Pro with bus lanes that are much faster and more abundant.

Yes, Apple is *the* customer for Thunderbolt. PC users might need certain PCIe based add-ons, but they don't necessarily need Thunderbolt. Apple decided to exclusively pursue form factors which preclude conventional user-accessible PCIe expansion methods, leaving Thunderbolt as the only PCIe expansion option on Macs for the past several years now. Notably, this transition also coincided with the availability of USB 3.0 on all new platforms, which, as you keep pointing out, eliminates the need for more costly PCIe based expansion solutions for the majority of people.

First you includes the entire mobile market in your sentence and THEN attribute it to only 5% Mac sales. WTF?!? Stick to one idea at a time. Apple controls less than 20% of the mobile market and less than 10% of the desktop market. PROFIT has NOTHING to do with adoption rates *OR* USAGE of a particular technology. Who cares how much money Apple is making in regards to Thunderbolt ADOPTION rates??? They have NOTHING and I mean NOTHING to do with one another. Your logic makes no sense AT ALL. Give it up. You can't make even a fair argument, let alone a good one.

You parsed my statement incorrectly. In each of those three markets taken separately, Apple has demonstrated the ability to reap the lion's share of the profit without necessarily having a majority marketshare. The most glaring example being the PC space, where in Q4 2012 they controlled only 5% of the global market yet took in 45% of the total operating profit—and we're just talking about Macs here. My argument, which is 100% sound, is that concentrating on selling fewer units but with higher ASP's and higher margins can be quite profitable. You keep arguing that Thunderbolt is dead because adoption rates are too low and the devices are too expensive for most people... Well as long as those devices are profitable for the parties that sell them, why would they ever go away? That's just economic reality.

Ah, so you're losing the argument by leaps and bounds and you then drop the "Troll" card. This is where I stop talking to you because if you've lost the argument 100%. I am not a troll. I've been on here for almost a decade and own four Macs. How the hell am I troll? Because I don't agree with your opinions? That seems to be the most common argument these days for "trolling". They don't agree with me so they must be a troll. How can they not see things just like me? I hate everyone that's different from me. Yeah, that pretty much sums up ALL conflict in the world in its entirety save that caused by greed including all political, religious and taste arguments.

Sometimes I seriously have trouble telling whether I'm engaged in discourse with a particularly dogmatic individual or they're simply yanking my chain. I didn't say you were a troll; it was merely a backhanded attempt at trying to persuade you to be a bit less imperious.

What dependency? Apple already has USB 3.0 in its entire lineup now. They can add USB 3.1 quite easily. And if they really cared about PCI expansion, they should add an actual professional tower back into the mix instead of that freaking trash can look-alike that SHOULD have been the next Mac Mini or xMac instead of a replacement for a Mac Pro which it in NO WAY functionally replaces (and requires a load of outboard awful looking mess of equipment to even attempt to replace, assuming one doesn't need the higher speed PCI lanes. How is THAT better than a tower that lets you keep everything internal and neat? It's not. It's a freaking GIMMICK that really only works for video editors and very few others without a messy desktop.

So to answer your question, bring back REAL PCI expansion instead of a R2D2 look-alike where even the video cards use custom connectors so you can't even replace the video card without paying the Apple Tax.

Who asked for Apple to make a Mac Pro without PCI expansion? NO ONE!

How can Apple quite easily add USB 3.1 to their Macs? The answer, of course, is that they can't until Cannonlake when Intel adds it to the chipset. Do you even think about this stuff before you rant about it? So 2017 is the soonest Apple might be able to drop Thunderbolt is what you're saying? Or do you expect Apple to announce that the compact cylinder form-factor experiment was financially unsuccessful after just a single generation and go back to doing a slightly more refined version of what every other OEM in the industry offers for a Xeon workstation?

BTW, I'm certainly not trying to change your opinion about what PC is right for you, or what designs might best fit your workflow. If anything I'm simply trying to temper your frustration by framing the situation in terms of the engineering and economics that shape the industry responsible for these products. Apple makes very few products for a company their size. If the new Mac Pro design bests that of its predecessor in terms of sales volume, margin or both, then everything else ends up being largely irrelevant.
 

MagnusVonMagnum

macrumors 603
Jun 18, 2007
5,193
1,442
How can Apple quite easily add USB 3.1 to their Macs? The answer, of course, is that they can't until Cannonlake when Intel adds it to the chipset.

Nonsense. They COULD add it as soon as a chipset is available. Just because they waited on USB3 only says they waited too long to put it in their computers. They made room for Thunderbolt. Remove it and you have plenty of space. Macs are getting too thin at the cost of other features and the phones are starting to show potential for being too thin to the point of bending in some cases. It's time to stop the obsession with thin and just be reasonable. I'd rather have a better computer that's a bit thicker than a thinner one that's crap.

Do you even think about this stuff before you rant about it?

It's quite obvious to me that you don't (see above). Integrated motherboards will not be the first 3.1 chipsets on the market. It may even appear before the end of the year according to some estimates. Why would Apple need to limit themselves to integrated motherboards sets? They're the richest company in the world. Waiting for everyone to pass them by is not the best business decision in the world, IMO.
 

toke lahti

macrumors 68040
Apr 23, 2007
3,270
502
Helsinki, Finland
Apple rarely breaks out sales figures by individual product, but I'd be curious as to how the Late 2013 Mac Pro has performed sales-wise compared to any of Apple's previous workstation class Macs.
Now it would also be good time, for even some 3rd pary, to make a survey among nMP owners, how much would they like to have upgradeability to 5k displays and usb3.1+DAM without replacing (=selling old + buying new) the present machine.
Or do you expect Apple to announce that the compact cylinder form-factor experiment was financially unsuccessful after just a single generation and go back to doing a slightly more refined version of what every other OEM in the industry offers for a Xeon workstation?
What many of us still wish that Apple could introduce real headless desktop computer, like G4 once was. Now they have funny Xeon workstation and über-tiny small factor machine. If there were expandable desktop machine in addition to nMP, people could vote with their wallet. Since macs are economically just a small hobby and ancient heritage for Apple nowadays, they shouldn't have any problems with this.
 

logician

macrumors newbie
Nov 29, 2015
2
0
I called this a couple of years ago. No one gives a crap about APPLE standards and Intel or not, it's ONLY been really used by Apple. You CANNOT create "standards" when you represent 8% of the market. Making matters worse, Apple wanted to have exclusive access to Thunderbolt the first year. WHY???? There was NOTHING available for it and it only meant that PC adoption would not occur for at least another year to even start. They helped doom their own connection standard. Call it Steve's dying folly if you like, but it was a stupid move.

First, as you note, it's an Intel-created standard, just like USB was. Apple didn't create USB but they did popularize it. (Can you name a single PC, motherboard, or peripheral that supported it before the iMac?)

Second, Apple's market share was even lower than 8% when the iMac put USB on the map. Market share can't be a prerequisite for popularizing a standard.

Your reasoning is backwards, which is why you have to stretch it in funny ways to try to make it fit the history. You make it sound like Apple's goal is to create standards, and usually they fail. Nothing could be further from the truth.

Apple (post-1997) hates creating standards. They learned that the hard way, from their first 20 years. They use whatever's out there, as long as it's good enough for them. (This is similar to what Seymour Cray did.)

They needed a good cheap pluggable interconnect for the iMac to replace ADB, and USB was good enough, so they used it. When they were looking into networking solutions in the late 1990's, WiFi was not yet a clear winner (remember those days?), but Apple found it to be good enough so they put their weight behind it. Just as the iMac was the first major computer to have USB, the iBook was the first major computer to have built-in wireless.

When Apple needed a fast pluggable interconnect in the 1990's, they found (but did not create) IEEE 1394, and rebranded it as Firewire. When USB finally caught up, Apple dropped Firewire. This wasn't a failure, any more than discontinuing Rosetta was, or killing the LaserWriter. It had served its purpose. Video editing and Firewire cameras were the desktop publishing and laser printers of the 2000's.

When Apple needed a fast flexible thin connection for iPods in the 2000's, they created the 30-pin connector. Then when they needed something smaller, they created the Lightning connector; Micro-USB lacked the features they needed (reversibility, physical robustness). Today, the new USB-C connector has finally caught up, so they'll eventually replace their custom connector with that.

You can see this with software, too. They dropped AppleTalk in favor of TCP/IP, and created ZeroConf (over IP) to implement the remaining pieces that they needed, and made it an open standard. They dropped MacRoman. They added Git support to Xcode. They changed NIBs to XML. They've got Secret Sauce, but they want as much of their infrastructure to be bone-standard as possible.

If you assume that Apple's goal is to create proprietary standards, you'll have to squint a lot to make the facts fit this theory -- Apple *technically* didn't invent that, and their market share was *so-and-so* back then! If you work from the theory that Apple hates creating standards and only does that when the industry-standard solution isn't up to par, then you can understand why and when Apple creates their own, and when they'll phase them out.

You see Apple standards as them saying "Hey everybody, here's a new way, please use our way, please?" In fact, it's more like "None of you guys could figure out a half-decent way to do this, so we had to show you. Here, make something like this. (And when you get your heads out of your butts in 5 years, we'll switch, but in the meantime, we'll enjoy our 5-year head start.)"
 

MagnusVonMagnum

macrumors 603
Jun 18, 2007
5,193
1,442
First, as you note, it's an Intel-created standard, just like USB was. Apple didn't create USB but they did popularize it. (Can you name a single PC, motherboard, or peripheral that supported it before the iMac?)

Second, Apple's market share was even lower than 8% when the iMac put USB on the map. Market share can't be a prerequisite for popularizing a standard.

I see it's time to resurrect the dead threads time. What, do you just search my old posts now and try to start new arguments? It sure as hell looks that way to me. :confused:

Just because you added USB first does not in ANY WAY SHAPE OR FORM mean you POPULARIZED it. The two don't add up! If you're 4-6% of the market (at the time), how the hell do you "popularize" something? If you look at actual history, Apple AVOIDED USB 2.0 when it came out. They PUSHED Firewire 400 instead even as far as making it a requirement on their 1st generation iPod (the one thing Apple DID "popularize"). This was a horrible mistake (fortunately, PowerMacs could easily add USB 2.0 with a simple card unlike today's Macs that can't add a damn thing without some massively expensive external thunderbolt rig) that even Apple realize after a year of complaints (how the hell do you get Firewire 400 on a PC when 95% of them never included it back then and probably around 75% never EVER included it in the history of Firewire?) Apple soon caved and added USB 2.x to Macs and switched iPods from Firewire to USB. We all know that USB 2.0 was inferior to Firewire 400, but it's precisely because Apple CANNOT "popularize" something with a small market share that their bid to make Firewire 400 the next big thing failed.

Your version of history has a 4-7% market share "popularizing" USB 1.0. Apple did switch to USB 1.0 early. They avoided USB 2.0 when it came out for over a year (and again with USB 3.0) so your argument falls apart. Do you seriously think PCs adopted USB because a couple of million (out of 300-400 Million at the time PCs total) Macs had it? Did the PC market adopt Firewire because ALL Macs had it? No. PCs never cared much for Firewire. Did PCs adopt AppleTalk? No. Did PCs adopt PowerPC? No. Did many PCs adopt Thunderbolt even though Apple is way more popular in general now than 20 years ago? NO, NO and still NO.

You have to have a significant market share to "popularize" something. The iPhone is a significant market share. iPhone and iPad changes start trends. Even the Apple Watch has a chance to popularize because it's a new slow to take off market and people are watching Apple now. But is Apple making Lightning available to others? Is it doing something new or just repackaging something in an EXPENSIVE proprietary connector like Apple seems to like to do (e.g. iPhone's entire history)? No one else asked to adopt Apple connection standards because they'd charge too much if they'd even do it (they like charging their customers exorbitant amounts for dongles and cables). But no one was really watching what Apple back in 1996 except to watch them self destruct and nearly go bankrupt which is exactly where they were quickly heading at the time. Yeah, what an excellent model to copy. A company that's going bankrupt because they lost the PC wars. Let's get us some of THAT! NOT! :rolleyes:

Did you know that Commodore sold more C64s that were more or less feature identical to each other in ten years from its release in August of 1982 than Apple sold in Macs (by a factor of 3) up until the point Steve Jobs returned? Apple sold 20 million Macs by 1997 since 1984. Commodore sold over 60 million C64s by 1992 and unlike the Mac, they were the same computer made in 1982 save cost saving motherboard changes and the like (different case design and color). And yet Commodore still died because they could not adapt the Amiga to confront the juggernaut that was the PC. That same juggernaut would have killed Apple too if Steve hadn't come back when he did.

Your reasoning is backwards, which is why you have to stretch it in funny ways to try to make it fit the history.

There's a big difference between Thunderbolt and something like AppleTalk. Thunderbolt was created to be a STANDARD and despite you trying to say it's all Intel's baby, you must have missed the original articles that stated Apple was working WITH Intel to create Thunderbolt and that's a big change in Apple's role of "creating" something in that Thunderbolt was supposed to become a standard. Apple preferred the superior Firewire 400 over USB 2.0 and rightly so (believe it or not Jobs was interested in standards when he came back; every move he made pushed the Mac away from its proprietary crap and towards standards or creating new standards including the move to Intel from PowerPC.) But Apple played a part in Thunderbolt's initial development and that's the key difference you seem to be missing when you want to paint Thunderbolt as just another Firewire adoption.

Meanwhile, Lightning is a joke. It was created just on the cusp of USB-C appearing that would have solved the problem and met a standard AND have had USB 3.x speeds from the start. Now Apple can't admit that it was a mistake despite the technical advantage of USB-C over Lightning. Arguing Lightning is slightly smaller is a weak argument when most people don't want phones to get any thinner, especially if it means sacrificing battery life. Lightning's ONLY advantages over USB 2.x is the reversible connector (and some creative types made a spring loaded backwards compatible reversible connector even there). Other than the new iPad Pro, Lightning is still USB 2.x speeds and I find it sad that Microsoft's new Lumia 950 series phone has faster wired connections than the latest iPhones and faster charging as a result of using USB-C. It also has wireless charging (something Apple fanboys will claim Apple "invented" on the next iPhone or maybe the one after when Android has had it available for years).

What I wrote that you quoted was over a year ago and as you can see by now that Thunderbolt has made a huge change in order to try to save/re-invent itself by moving to USB-C connectors (when the USB guys didn't even want Thunderbolt to use the same connector from day one) but somehow made it "OK" by including USB 3.x within the Thunderbolt standard itself as a subset controller (even though that doesn't change the confusion factor when wondering why a Thunderbolt 3 device doesn't work when plugged into a PC that only has USB-C). Thunderbolt is desperately trying to stay relevant in a world that doesn't really need or want it. Still, this is a good move for it for now, but the problem is it MUST stay well ahead of all future USB changes to in order to keep this ruse of "1 connector to rule them all" relevant. If USB4 got a jump on Thunderbolt 4 and was faster they'd have a real problem on their hands and Thunderbolt would be moot once again. Nothing changes the fact that Thunderbolt connections (despite the same connector cord) require more hardware to function at those speeds and thus cost more for the same device. A 6TB conventional "media" hard drive capable of mere ~180MB/sec speeds at maximum is not going to be ANY faster with a Thunderbolt controller than a USB 3.1 controller and thus it makes ZERO sense to use a Thunderbolt version that will inevitably cost $200 more than the USB 3.1 version, ESPECIALLY if they use the same freaking port on your next generation computer! Hey, charge me $200 for NOTHING!

I see ONE major advantage of Thunderbolt III that might make it viable even so and that is external graphics card hubs that would essentially let you do ONE WIRE to connect and turn your notebook into a gaming or graphics workstation capable desktop when docked (yeah you "can" do this with Thunderbolt II, but it's too slow and isn't directly designed around this which means it costs more whereas Thunderbolt III has this in mind from day one). Given desktops are already a dying breed compared to mobile platforms, this would indeed be a useful trick to spare someone from having to own a second entire computer just for home/game use, but ONLY if said dock costs significantly less than an entire second computer and sadly, that hasn't really been the case in the PC world where $600 + a graphics card is insane when you can build a gaming computer for $1200 no problem (less if you don't need 4K gaming). Mac users would crap their pants for one, though since there is no gaming Mac at any cost.


I don't know how you come to that conclusion.
Apple (post-1997) hates creating standards. They learned that the hard way, from their first 20 years. They use whatever's out there, as long as it's good enough for them. (This is similar to what Seymour Cray did.)

I think you mean pre-1997 and post 2011 as Jobs clearly pushed for standards, moving the Mac from an obscure IBM processor to the world's x86 home standard and making Mac software easier to port, making sure Macs had standard PCI buses on the Pro "trucks" (broken by Jonny Ive with his custom GPU connectors and no expansion slots), supporting OpenGL (standard) instead of licensing DirectX (which I'm sure Microsoft would have licensed them at a price, especially given they needed the Mac to pretend they weren't a total monopoly), standard gigabit Ethernet at a time when most PCs only had 100Mbit, adding HDMI even to notebooks (now largely gone again) etc. etc. Steve's Apple only went proprietary when necessary, not just to make an extra buck on sucker users. That's the OLD Apple that wasn't under Steve's control and sadly it appears to be the new Apple under Tim Cook's control.

They needed a good cheap pluggable interconnect for the iMac to replace ADB, and USB was good enough, so they used it.

Don't you mean that Apple was in dire straits at this point and couldn't afford to roll their own incompatible standard? ;)

I maintain the old non-Jobs Apple would have gladly sacrificed compatibility with the IBM PC clone market if it means they could SOAK a large enough Mac market for overpriced proprietary hardware. They are still soaking the market today with dongles that cost 2-4x what they should cost and released at times where it's questionable to do so (i.e. no one is using mini-display port, so shouldn't it at least come with a free dongle? No, why give away for free what you can charge over $20 for a $1 part and make $19 profit? Hey, the new post-Jobs Apple isn't so different, after all. :eek:

When they were looking into networking solutions in the late 1990's, WiFi was not yet a clear winner (remember those days?), but Apple found it to be good enough so they put their weight behind it.

Apple was also nearly bankrupt at the time as I've mentioned so there was no capital to invent something new. Besides, as I've indicated once Steve returned circa 1997, it was more like NeXT than Apple at that point (you know the Apple that put Steve in virtual "Siberia" in 1985 and caused him to leave Apple since they would not let him work on the Macintosh any longer).

Just as the iMac was the first major computer to have USB, the iBook was the first major computer to have built-in wireless.

And again, being one of the first, doesn't mean you "popularize" something. Major computer? In what sense were the Powerbooks ever "major" computers anyway? They were a blip on the radar at most.

When Apple needed a fast pluggable interconnect in the 1990's, they found (but did not create) IEEE 1394, and rebranded it as Firewire. When USB finally caught up, Apple dropped Firewire.

You mean a few years ago with USB 3.0? Apple never "dropped" Firewire even long after it was an obvious commercial "failure" and some would say that was a good move since professionals used Firewire even if most regular users did not. Even so, dropping it has had consequences for many commercial audio rigs that still use Firewire (somehow adapters never work quite as well).


Regardless, I don't know WTF you're trying to argue, really. Your entire post is predicated on the words helped "create" vs "adopt" (despite Apple's obvious hand in Thunderbolt's development; they're also on the USB board, BTW so they do have a "hand" in USB standards as well) and trying to change the REASONS Apple did what it did from "greed" to "convenience". Otherwise, you're preaching to the choir.

When Apple needed a fast flexible thin connection for iPods in the 2000's, they created the 30-pin connector. Then when they needed something smaller, they created the Lightning connector; Micro-USB lacked the features they needed (reversibility, physical robustness). Today, the new USB-C connector has finally caught up, so they'll eventually replace their custom connector with that.

You went from arguing that Lightning was better than USB-C (being slightly smaller than it) in the other thread to "eventually replace" here. o_O That's like saying I was right all along but not admitting it and then starting to argue here about something I said over a year ago in the context of "they don't create standards; they adopt them!" when clearly modern Apple does help create standards and they conveniently switch to whatever sells them high-priced adapters and dongles as well. Apple loves to over charge still. The only difference is that for awhile there, the higher prices bought you premium features (there was no better notebook for the dollar in 2008 than my Macbook Pro, even for Windows it ran Vista better than the rest of the PC world). Now it just buys thinner and thinner and thinner instead of better and better and better and sadly Steve's own gauntness seemed like it mirrored his last few years on earth obsession with "thin" and Apple is still obsessed with it even when it means their GPUs are pitiful compared to machines that cost half as much.

If you assume that Apple's goal is to create proprietary standards

It's funny, but I don't recall saying that about Modern Apple. The OLD Apple used to do that (well I wouldn't call them "standards" exactly since only Apple ever really used them) in the '80s and early '90s (got to be different) and it nearly bankrupted them. The Tim Cook Apple had better be careful what it does because history has a way of repeating itself. For a company that touts the latest Intel CPUs all the time, they ignored USB 3.0 for a long time while throwing Thunderbolt out there when it wasn't useful to anyone (nothing available for it) and mini-display port when no one had it (and display port hadn't taken off) leading to dongle land once again.

NO ONE ELSE has adopted Thunderbolt + Mini-Display port in any quantity and now it's moving to USB-C connectors and Apple seems to be sitting on the edge of the fence trying to make up its mind whether it wants to support that standard or not. They put it on ONE 13 inch Macbook and NOTHING else since then (and there's been a whole new release cycle since then) and even the model they have it on has only ONE port (including charging) so it's a rather poor POOR implementation requiring (you guessed it) more adapters and hubs. They SHOULD have put it on ALL the rest of the Macs as an extra port for now until it does catch on, giving flexibility and all larger models should have at LEAST two ports, preferably more since it does everything all the other ports put together did.

If Apple doesn't put Thunderbolt III in a USB-C package next spring, then they are stupid and the Mac deserves to die. This is their one chance to get out there early with a notebook that can one wire connect to an external graphics card + hub AND maintain the new standard that will be EVERYWHERE within three years. Wait on this and I know I'm moving on. I've been holding off on my next notebook AND desktop purchase because of the Thunderbolt III + USB-C promise of combination computer/dock setup that will make my life simpler. But I'm not going to wait forever when PCs will be all over this like glue.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.