Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Thunderbolt Adoption Reportedly Slowed by Intel's Licensing and Certification, But Improvements Coming
- Thread starter MacRumors
- Start date
- Sort by reaction score
Are you serious?
A copper Cat6 cable with GbE is far faster than WiFi, yet still fails to keep up with current spinning hard drives.
The skeptics here complain that USB 3.0 is only three times faster than Cu GbE....
Without trying to be 'smart' here, there is a $29 solution for that. It would have been more considerate of APPLE to also include legacy ports, during the inevitable 'transitional' period though.
My point was not that WiFi is faster. But it wins on convenience, which is more important for people in general from my experience.
Those who need cables are mostly power users.
For every poster who has declared Thunderbolt dead, are you so sure that you'd be willing to donate one penny for every Thunderbolt controller Intel sells in 2013 to a cause of my choosing? How about in 2014 then? 2015 maybe?
I'm not detecting sarcasm here, so I'm just going to assume you're WAY off with your numbers. Just to be clear, there are several layers to many of these licensing structures. You have developer program fees, logo licensing, device certification and per unit (or even per port in some cases) licensing fees. That's all separate from the cost of the controller itself, which is generally the most expensive part of the equation.
So a top of the line DSL3510L 4-channel Cactus Ridge Thunderbolt controller currently runs about $30. USB is a little different because the device silicon is different depending on the device class, but a typical SuperSpeed USB to SATA 3 Gb/s bridge chip does run about $5. In either case, the licensing costs per unit (if you actually sell more than 1000) are going to be less than the controller costs.
There is no doubt that designing, engineering, writing drivers for, validating, and obtaining certification for a Thunderbolt device is a daunting and possibly rather expensive ordeal. I think this is why a lot of OEMs are just saying, "Intel's licensing is too expensive," rather than doing the hard work.
edit: I also forgot to point out to all those asking for Thunderbolt for their iOS devices that a 1st gen Thunderbolt cable uses more power than the SoC in an iPhone 5, and a 2-channel Cactus Ridge controller uses almost as much again while being almost 50% larger than the A6. Which leads to the other item people seem to think might happen, which is integration with the chipset. The 4-channel Cactus Ridge controller is just about as big as the PCH and costs almost as much, so I don't see these two getting hitched. Unless Intel switches from FDI and DMI to using Thunderbolt as the protocol for the link between the PCH and the CPU...
I'm not detecting sarcasm here, so I'm just going to assume you're WAY off with your numbers. Just to be clear, there are several layers to many of these licensing structures. You have developer program fees, logo licensing, device certification and per unit (or even per port in some cases) licensing fees. That's all separate from the cost of the controller itself, which is generally the most expensive part of the equation.
So a top of the line DSL3510L 4-channel Cactus Ridge Thunderbolt controller currently runs about $30. USB is a little different because the device silicon is different depending on the device class, but a typical SuperSpeed USB to SATA 3 Gb/s bridge chip does run about $5. In either case, the licensing costs per unit (if you actually sell more than 1000) are going to be less than the controller costs.
There is no doubt that designing, engineering, writing drivers for, validating, and obtaining certification for a Thunderbolt device is a daunting and possibly rather expensive ordeal. I think this is why a lot of OEMs are just saying, "Intel's licensing is too expensive," rather than doing the hard work.
edit: I also forgot to point out to all those asking for Thunderbolt for their iOS devices that a 1st gen Thunderbolt cable uses more power than the SoC in an iPhone 5, and a 2-channel Cactus Ridge controller uses almost as much again while being almost 50% larger than the A6. Which leads to the other item people seem to think might happen, which is integration with the chipset. The 4-channel Cactus Ridge controller is just about as big as the PCH and costs almost as much, so I don't see these two getting hitched. Unless Intel switches from FDI and DMI to using Thunderbolt as the protocol for the link between the PCH and the CPU...
Last edited:
Good luck connecting external monitors over USB 3.0 instead of Thunderbolt.
The world has been able to do this over USB 2.0 for quite some time already, so, of course, the greater bandwidth of USB 3.0 was hardly going to be a stumbling block.
http://www.amazon.co.uk/dp/B00612ZPQA
You can stop rolling those eyes now.
I hate coming into the middle of a conversation, since I have no idea if you're being sarcastic or not. But...
Every computer sold in within (at least) the last 5 years have PCI-e onboard.
That's cool! And almost the cost of a TB cable
The resolution support is a downer though and they seem to have performance issues. So it can't really be compared to the displayport.
In all the years of firewires existence, i hadnt used a single firewire device (in fact no PC i owned had firewire ports) until i was purchased by work, in late 2008, a macbook pro.. the only device i have ever connected to it, is the also supplied by work Drobo.
Likely the only device ever to connect to my iMacs thunderbolt port will be a USB 3 breakout box, and possibly a Drobo mini (because the idea of 4 raided 256gb ssds makes me very happy)
This is all also keeping in mind that the price gap between firewire products and USB products became negligible very quickly, and a lot of devices carried both as standard, i don't see this happening with USB3/Thunderbolt because people will buy a much cheaper USB3 only competitor to a product with both ports.
Good luck connecting external monitors over USB 3.0 instead of Thunderbolt.
USB 3.0 + Thunderbolt is the perfect combination IMO. USB 3.0 for all the cheap consumer stuff, and Thunderbolt for the high-end stuff and monitors and docks.
Uh actually monitors over USB 3.0 will be fine. It was possible to do it over USB 2.0 however it was noticably slow. 3.0 has way more bandwidth, and can easily handle a display.
So
right back at ya!----------
That's cool! And almost the cost of a TB cable
The resolution support is a downer though and they seem to have performance issues. So it can't really be compared to the displayport.
I dont think you'd ever use it over thunderbolt/display port, however if you've only got 1 TB port and you use it for something else, going with USB 3.0 for a second screen isnt such a bad option. It'll be perfectly fine for a computer display.
Obviously you may have a few slowdowns/low refresh rate if you start doing full HD videos on it, but even then, it wouldn't be too bad.
Take a look at this video (note that its from 2010 and the hardware in this video is VERY bulky. These things are much smaller now)
http://www.youtube.com/watch?v=MBBDHH-8-yc
Uh actually monitors over USB 3.0 will be fine. It was possible to do it over USB 2.0 however it was noticably slow. 3.0 has way more bandwidth, and can easily handle a display.
So
----------
To be fair, most people won't use these nor are they ideal. The big problem with USB is that everything necessarily leverages the CPU to do the grunt-work, and this kind of a device offloads what the GPU normally would do onto the CPU. Objectively, from a technical standpoint, it is easily an inferior protocol.
That said, right now, it's hard to see the advantage of Thunderbolt when cost is figured into the equation. Everything you can do with TB, you can do with USB3, albeit less elegantly and with a bit less performance (but again, it's cheaper). It's just the reality of Thunderbolt adoption right now. The longer it takes to be price-competitive, the harder time it will have garnering market share. And this is coming from a TB proponent (and owner of both an 27" ATD and a TB-enabled Drobo).
Well, you could also daisy chain two TB displays.
But I can see using something like that for a smaller older monitor though, provided that it doesn't put too much load on the CPU.
TB feels much more future proof in this area anyways.
Well, you could also daisy chain two TB displays.
But I can see using something like that for a smaller older monitor though, provided that it doesn't put too much load on the CPU.
TB feels much more future proof in this area anyways.
I dont think 'smaller old monitors' come into it. Every single monitor that isnt an Apple Thunderbolt display comes into it.
Lets say you've got a thunderbolt hard drive and a macbook air. A USB 3.0 external display is the only option there, that or spending a hell of a lot of money on an out of date thunderbolt display.
Alternatively, if you want a second screen (and again, dont want to spend stupid amounts of money on buying a new screen 'just' for thunderbolt) its a great option.
I for instance have a rMBP 15". Both thunderbolt ports are used for my dual 24" monitors. I couldnt use the HDMI port as OSX doesnt support many HDMI monitors, and treats the connection to my monitor as a TV if I use that port (thus giving overscan and terrible picture quality).
I would however consider a USB 3.0 graphics card for one of the monitors if I could get one that plays well with a USB 3.0 hub (again, Apple screwed up the USB 3.0 implementation so most hubs dont work).
----------
No. Just no.
USB is highly relevant.
If you live in the middle of a field with nothing around you, wireless is great.
If you live in a city, its crap. I live in a fairly small neighbourhood, and right now I can pick up over 30 wireless points, covering all wireless channels.
Wireless performance is absolutely crap if all channels are in use.
----------
Which doesn't play well with a lot of monitors.
Actually, there's a wireless standard for peripheral connectivity that is about to be introduced, it operates on the 60 Ghz band, which unless you have multiple APs in a very close range (like less than the range between your chair and desk), won't be a problem.
Look up 802.11ad (not ac, ad)., also known as WiGig, capable of 7 Gbps speeds for peripherals interconnect without wires. This could replace USB (though Bluetooth is more susceptible to get replaced by this).
Except that those adaptors doesn't provide high enough resolutions for every single monitor. That might change though.
It's too bad you can't daisy chain from the ATD to a DP monitor.
People in general do not worry that much about performance from my experience.
Actually, I have never heard anyone besides power users complain about the speed of USB2.
It's good enough for what most people use it for, so is wireless.
And as KnightWRX pointed out, there is better tech coming up.
What do people use USB for?
HID - Doesn't require USB3
Cameras - More and more are using their mobile phones which will sync wirelessly. Pocket cameras are coming out with WiFi these days, even DSLR has the option.
Mass Storage - Network storage and the cloud is the future for storage and backups here since more and more use mobile devices.
Printers - People generally prefer network printers these days.
Scanners - Could be one that doesn't go wireless, unless it's built into a printer.
Anything else?
Like your laptop sitting on your desk next to your desktop? When the personal computers ( perhaps including tablet form factor also) are the access points then the number of AP will very significantly increase.
However, like bluetooth there will likely be pairing and other mechanism to somewhat ignore other broadcasters in close proximity. If it is a close proximity radio solution then things typically will get crowded.
Doubtful this will reach the low power and low implementation costs that bluetooth has. Similar to generic USB versus Thunderbolt what does the speed increase do for a keyboard, mouse, trackpad , etc ? Nothing.
For USB it is a similar issues. Plus for USB, Webcams , inexpensive flash drives , HDDs ? hardly nothing. Plus it provides power for those kinds of devices.
The bigger threat is probably more so aimed at wired cable of the nature of HDMI , DVI , DisplayPort. Killing that cable has so far faced much resistance. base computer box / set-up box / dvr box in close proximity of a monitor will tolerate the short distances.
Intel's WiDi http://en.wikipedia.org/wiki/WiDi tries to take on HDMI but without the bandwidth currently (since leveraging WiFi ).
That seems to be a bit of a contradiction, though. The REASON the prices are high is that no one is adopting it or buying the products (Chicken/Egg). It didn't help that Intel made it exclusive to Apple to start, though. That just delayed things and kept prices even higher. If you want something to be a STANDARD, you have to make it widely available and competitive. If Intel had to eat some costs to keep prices down until it took off, well it's been done by the printer industry for ages.
Yeah, but there's NOTHING "Convenient" about daisy-chaining. Just ask anyone on here that has ONE Thunderbolt port and it doubles as their 2nd monitor output. If they want to hook up a hard drive to it and a monitor is already connected, they have to disconnect the monitor, plug in the hard drive and plug the monitor into the pass-through on the hard drive (assuming it even has one). Then, the process has to be reversed when the portable hard drive is disconnected and that means fiddling with more wires. Now imagine having 3 hard drives connected and removing the one in the middle. It's time to look for the wires again. It's a MESS. It would be FAR better to have a HUB to just connect and disconnect to. And people pointed out there would be hubs for TB, but I haven't seen a single one yet that was more than just a bunch of OTHER ports and a single pass-through for TB and even most of those haven't actually made it to market yet over two years later. Apple's own monitor has USB 2.0 on it (i.e. It's fun to replace an entire overpriced monitor just to update its hub).
My Mac Mini should have at least two TB ports on it given it uses one as its second video output to at least partially avoid that mess, but they didn't bother and now I'd loathe to have temporary drives connected seeing I am using that second monitor connection.
Eventually (and that is the key word here), a few TB ports on each desktop or laptop could have replaced most of that other myriad of connections now found on almost all computers. A shame, really.
It's just not realistic with daisy-chaining. You'd still need a central hub to make that work without the daisy-chaining nightmare mentioned above. And obviously these hubs aren't going to be cheap any time soon. In short, TB is like a Porsche 911 Turbo. It's a GREAT car, but it's not without its drawbacks in high maintenance and upfront costs, making it largely unaffordable to the masses and it'll NEVER be cheap or widely adopted. It's a niche. TB is a niche. It's just too darn expensive when a Chevy Camaro costs less than a third as much and with 75MPH speed limits, the difference in practical daily driving is smaller than it could be if maxed out.
After over three years, I'm still hard pressed to find very many monitors that even support Mini-Display Port directly. Mine are HDMI and DVI and thus I need adapters for both.
That's actually too slow to be useful for things like high-end video cards outside the case and even a lower grade one would use all the bandwidth up. Thus, the ideas of having PCI external falls short. It's nice to say you have 10Gbps NOW, but if there are only a couple of Pro uber-expensive devices that can make use of that speed, it's just a niche. USB3 is already prepping for 2-3 years from now when consumer hard drives are expected to start catching up. The difference is that USB 3.1 will be dirt cheap at that time. TB will STILL be expensive (because no one is using it).
I consider this a negative. Daisy-chaining is a PITA to add/remove devices anywhere except the end of the chain and yet its presence discourages the creation of affordable hubs (if/when they do appear, they will NOT be affordable).
Yes, replace 50 cent connectors with ones that cost $30. That's totally worth it. NOT.
And just imagine trying to disconnect your mouse or keyboard in the middle of 4 hard drives and a sound card and a printer and then re-connect the remaining devices and then your friend wants his portable hard drive he brought back and do it again. Can you say NIGHTMARE? That's daisy-chaining for you. I have YET to see a single multi-port hub for TB so the possibility of it is moot to me. But once you admit that USB is more practical for mice, keyboards, etc., then you might as well admit that USB 3.0 is more practical to since it's dirt cheap and backwards compatible with all that stuff and cheap hubs abound.
At what point? Existing computers with TB will not be forward compatible so you'll need new hardware. By then USB 4 might be out. In other words, it's just sheer speculation about if/when a 100Gbps or higher device will be available and whether it will be competitive at that point in time. I don't bet on dreams.
And cows can fly with a good deal of help. That doesn't make them competitive against birds.
----------
Wireless is SLOW and prone to interference. All three of my computers in my den are connected by Gigabit Ethernet. It's been around for over a decade and not even 802.11AC can beat it for low latency/overhead and usable speed. USB3 is up to 5x faster than Gigabit Ethernet. TB is 10x faster. I fail to see how Wireless is good for anything but portable devices that are often too slow to render a web page quickly so absolute speed doesn't matter a whole lot.
----------
Get one of the $3 (with free shipping) adapters off somewhere like Amazon like I got. My TB port is happily driving a 2nd monitor with HDMI right now.
If you read my other posts you will see that I'm talking about the average consumer and I think that the convenience is more important than the performance, provided it's good enough. People LOVE wireless.
Power users will still need the cables though for reliable performance.
All I'm saying is that wireless in my den would be utterly pointless since the desktops there aren't going anywhere and so it's a one time cable connection and then forget about it. Using 802.11N would only degrade the network connection, consumer or not. Obviously, wireless has its uses (single notebook only used for Internet really wouldn't notice a difference with the average Internet connection typically being under about 20MBps still.
Now I would agree it's a lot more work to run Ethernet around the entire house if it's not built with it already. I am using wireless for my AppleTVs on the other side and even floor of the house, but I have to say that I still get glitches sometimes and wish I had bothered to run the Ethernet (and still might if I get some time to do it).