POST A COMMENT

44 Comments

Back to Article

  • Kevin G - Monday, April 08, 2013 - link

    DP 1.2 support? I see this as being the gateway for Apple to finally update the Mac Pro line. They've been holding out on the Mac Pro update partially due how to implement Thunderbolt on in their professional line up. Now they have the means to get 4K output out of a single connector. I figure these controllers will be on the GPU PCB as special edition Geforce or Radeon card specific to the Mac Pro's. Reply
  • FeelLicks - Monday, April 08, 2013 - link

    I wonder if DP 1.2 support is the cause for delay of Belkin's TB Express Dock? Or is it too late for that? In which case it makes buying the Dock a very UN-future proof purchase at this time... Reply
  • repoman27 - Monday, April 08, 2013 - link

    The first and second gen Thunderbolt controllers (Light/Eagle/Cactus Ridge) were designed specifically to accompany LGA 1155 Sandy Bridge and Ivy Bridge systems. Mac Pros are Xeon based. The lack of an LGA 2011 Xeon with integrated graphics or some other graceful means of piping the DisplayPort output from the dGPU to the Thunderbolt controller is more the issue at hand. Until this past December, DisplayPort 1.2 displays weren't even available, so it's not like Thunderbolt being limited to DP 1.1a would have held the Mac Pro refresh back in any way.

    The DP 1.2 support would not be relevant to Belkin's dock, it's only useable by the host PC. There may well be some other items around the edges that have now been integrated and which make waiting for Redwood Ridge desirable, but I think Belkin's issue is more to do with drivers and engineering.
    Reply
  • Kevin G - Tuesday, April 09, 2013 - link

    Thunderbolt controllers are somewhat generic. However, a Thunderbolt add-in card for PC's hasn't reached the market for the same reasons why it hasn't appeared on the Mac Pro: tunneling video through it. Outside of socket LGA 1155 motherboards which have integrated graphics to funnel, the only other logical spot would be on a discrete video card. Neither nVidia or AMD wish to support this standard. In the case of Apple, they'd have to design their own discrete card for this purpose.

    The case for DP1.2 is simple: Apple is pushing higher resolutions and that means 4K in the professional space. Apple did this previously when they released the 30" CinemaDisplay and were the first major push for dual link DVI in the market with a special Geforce 6800 Ultra card (PC versions were initially single link DVI). Apple wants the support to be there even if most customers will reuse their existing displays with it.
    Reply
  • Penti - Monday, April 08, 2013 - link

    Ivy Bridge-EP is probably the key here simply. Nothing else happens on the Xeon side. Haswell-based system would take a sidestep for it from being a workstation. Nehalem/Westmere is still pretty powerful.

    However it's pretty ridiculous that the Thunderbolt Display monitors from Apple actually don't support a pure DP input from a Mac Pro. Not sure how they will address it in the Mac Pro, but it will probably be pretty different from their basically 4 year old machine that they didn't even bother to validate it against the new IEC 60950-1 rules. Not sure they need to address that though, their customers don't sit on Thunderbolt Displays.
    Reply
  • DanNeely - Monday, April 08, 2013 - link

    If Intel wants TB to gain any traction in the PC market they're going to have to bundle TB into at least some of their higher end chipsets. Firewire struggled to gain and maintain traction when it was a $10 extra and had performance that crushed USB1 like a bug. For every consumer task except video out USB3 is good enough which puts TB in the place where FW was after USB2 came out and cheaper USB2 devices ran FW out of the market. Reply
  • jeffkibuule - Monday, April 08, 2013 - link

    Yep, it's pretty easy for OEMs to ignore an extra $20 part for a technology almost no one uses. If Intel wants widespread adoption, they need to stick it in the chipset, but that's not going to bring them significantly more money. Reply
  • royalcrown - Tuesday, April 09, 2013 - link

    UH.....you forgot that the manufacturer also need to quit charging 50 dollars for a 3 foot cable and there needs to be actual peripherals besides pre built external hard drives. If I could replace my OWC external drive with reasonably priced TB, I would ditch FW800 in a SECOND. Reply
  • DanNeely - Tuesday, April 09, 2013 - link

    With all the electronics needed in each end of a TB cable there's never going to be a $5 monoprice version. Meanwhile the chicken and egg problem is slowing what potential cost savings are available by replacing several generic transceiver chips with a single TB specific one. Reply
  • danbob999 - Tuesday, April 09, 2013 - link

    If intel integrated Thunderbolt into the chipset, it would be a great gift to AMD since it would drive up the price of their chipsets.
    Thunderbolt is the new Firewire. It will never get mass market adoption and will be phased out with the introduction of USB4.
    Reply
  • geniusloci - Tuesday, April 09, 2013 - link

    And where is USB4 going to be when TB is doing 100Gb/s ? Reply
  • mayankleoboy1 - Tuesday, April 09, 2013 - link

    Doing 200Gb/s.

    Why are you assuming that USB4 is going to have lesser bandwidth than TB ?
    Reply
  • web2dot0 - Wednesday, April 10, 2013 - link

    Right .... 200Gb/s without active components .... right. Talk alot of talk, but no actual evidence to back up their claim.

    Thunderbolt can easily get to 1Tb/s if they follow through with the technology evolution. it's simply not possible with USB without architecture change. That freaking $5 cable isn't going to get you there ....
    Reply
  • mayankleoboy1 - Wednesday, April 10, 2013 - link

    Do you even know that TB is built upon the PCIe lanes ?

    " That freaking $5 cable isn't going to get you there .... "
    But it will give 100W power, and 10GB/s bandwidth.
    Reply
  • daneren2005 - Wednesday, April 10, 2013 - link

    Doing the amount people actually need at the time, for the price people actually want to pay. Reply
  • xdrol - Monday, April 08, 2013 - link

    ".. official support for DisplayPort 1.2 (and thus 4K displays). If you connect either of these parts to a Thunderbolt display you still only get DP 1.1a support."

    This is confusing, when is this just 1.1a, and when is it supporting 1.2?
    Reply
  • FeelLicks - Monday, April 08, 2013 - link

    Yeah, this part was confusing to me too.

    My take:
    Cactus Ridge = 1.1a
    Redwood Ridge = 1.2
    Reply
  • Kevin G - Monday, April 08, 2013 - link

    I think that the reference is to Apple's ThunderBolt display, not the generic idea of a display connected via Thunderbolt.

    This certainly can use some clarity.
    Reply
  • FeelLicks - Monday, April 08, 2013 - link

    Yeah, I think you are right. It means even if you use Redwood Ridge (which supports 1.2) to connect to an older Thunderbolt display, since the old display will have Cactus Ridge (port) it will still be 1.1a. Reply
  • repoman27 - Monday, April 08, 2013 - link

    This is a bit confusing, but I think the deal is that Redwood Ridge supports DP 1.2 only when operating in DisplayPort signaling mode. The DisplayPort sink to Thunderbolt protocol adapters are probably still the same as previous generation controllers, since a 10 Gbit/s channel can't carry a full DP 1.2 HBR2 main link.

    If you plug a DP 1.2 panel directly into a Thunderbolt port on the host PC, you get DP 1.2 output. If you daisy-chain through another Thunderbolt device, you only get DP 1.1a.

    I am curious if MST might still be supported though, even if HBR2 isn't.
    Reply
  • ltcommanderdata - Monday, April 08, 2013 - link

    Since current Thunderbolt is PCIe 2.0 based is Thunderbolt 2014 PCIe 3.0 based to achieve the doubled bandwidth? Reply
  • JNo - Monday, April 08, 2013 - link

    "Redwood Ridge"

    Silly name. Say it fast as you can 10 times in a row - makes a great tongue twister though.
    :D
    Reply
  • HisDivineOrder - Monday, April 08, 2013 - link

    The reason only Apple "gets it" is because Apple didn't really emphasize USB3 very much. Meanwhile, PC's did. So when Thunderbolt shows up, is insanely expensive, offers little in the way of reason to exist to the majority of consumers, then honestly Intel is providing an answer to a problem that's already been solved. Namely, faster than USB 2 was needed and despite Intel's best delays to try and stymie support, USB3 was adopted. Thunderbolt remains outrageously expensive and peripherals that support it remain fringe expensive.

    So whose fault is it? Intel's. Completely and totally Intel's. Like with everything, Intel charges way too much for it and acts like they're ahead of everyone else, so why wouldn't people pay more for their tech? Except USB3 is perfectly fine for most anything people use the tech for. Most people don't need or want to pay for a slightly better DisplayPort cable when anything that includes the tech is so expensive as to make people's eyes bug out.
    Reply
  • Tegeril - Tuesday, April 09, 2013 - link

    You really have no idea what Thunderbolt is capable of if you think it's just a 'slightly better DisplayPort cable.' Reply
  • watersb - Tuesday, April 09, 2013 - link

    But the 'have no idea what Thunderbolt is' is *exactly* the huge problem here. Reply
  • mayankleoboy1 - Tuesday, April 09, 2013 - link

    This.
    TB is just a layer on top of the PCI-e lanes. It is not really revolutionary. It just uses more bandwidth to give more bandwidth to the end user. There is nothing special about it.
    Reply
  • philipma1957 - Monday, April 08, 2013 - link

    As a t-bolt user and owner of 1 pegasus r6 ,1 pegasus j4 , 3 lacie little big disks and 1 seagate stae 121 t-bolt gear is sooo much better then usb3 it is funny.

    The real problem is most people don't need it for what it can do best ; allow a very fast external boot setup. Most window pc's have access to the internal area and putting in a ssd as a boot drive is easy. right now iMacs are a joke to put in a ssd. It is hard to do a mac mini and apple ssd options cost an arm and a leg.
    so like it or not t-bolts as boot drives are mac-centric . If we get a 2x speed t-bolt it may allow external gpus so mac would benefit as upgrading your gpu is tough in a mac. (other then a mac pro) So i really don't see t-bolt working with windows gear unless most windows builders prevent easy hdd access. Just an opinion for what it is worth.
    Reply
  • name99 - Tuesday, April 09, 2013 - link

    I don't want to insult you, but I really think pretty much everything you said is false.

    (a) I would guess that most PCs these days are either laptops or iMac clones --- that's what I see selling in large numbers at Best Buy. So getting at drives is not easy on the bulk of machines.

    (b) Macs can (and have obviously for years) been able to boot off USB. I ran my old iMac for years off an SSD connected vis USB2. You can do the same thing (and a whole faster) via USB3, but even via USB2 if you really want, you could RAID stripe 2 or 3 SSDs together and use that as a faster drive. I've striped (using Apple's built-in SW RAID) many slow drives together in my time and it works pretty well.

    (c) I have no idea what the current Windows USB boot situation is. A quick Google search suggests to me that it's still pretty crappy.

    Point is, I don't think your logic works because actually all the evidence suggests that Apple users can externally boot off USB (2 or 3) just fine and can goose the performance if they want, whereas it is WINDOWS users who cannot easily boot off USB, and who would benefit from booting off TB (which I assume would look to the OS like booting off an internal drive).
    Reply
  • Dustin Sklavos - Monday, April 08, 2013 - link

    I'm kind of in the same boat. Thunderbolt is all potential and no execution. "Oh look, our refresh reduces power consumption and adds DisplayPort 1.2 support." And? That doesn't actually address any of the very real problems Thunderbolt has.

    The fact that the controller eats half the PCIe 2.0 lanes off of the PCH means that motherboard manufacturers have to either give up a lot of value adds or install a multiplexer, and neither of these is an attractive option.

    Bundle it with the chipset, make it cheap, and then maybe you'll see adoption. But expensive proprietary standards don't benefit anyone.
    Reply
  • repoman27 - Monday, April 08, 2013 - link

    That's a very odd way of looking at it.

    5% of all PCs shipped worldwide in the past two years have had Thunderbolt ports due solely to Apple's commitment to the technology. Thunderbolt is an Intel technology designed first and foremost for Apple, to solve Apple's problems, namely that scant few of the PCs they sell have user accessible PCIe slots or any available internal volume to house expansion.

    Thunderbolt is just the ticket for MacBooks, Mac minis, and iMacs, all of which have integrated graphics, and those that do have a dGPU generally have one that can get by just fine with only 8 PCIe lanes. Thus the most common configuration that Thunderbolt ships in is connected to the PEG lanes off of the CPU and not hung off of the PCH. The whole point of Thunderbolt is to essentially provide all the bandwidth of the PCH via one small friction-fit connector. The back end of a Thunderbolt controller is the equivalent of the PCH's DMI 2.0 and FDI connections.

    The fact that LGA 1155 doesn't offer enough PCIe lanes to go around is hardly a shortcoming of Thunderbolt. You could say the same of any bandwidth intensive controllers: 10GbE, 6 Gb/s SAS/SATA, USB 3.0, etc.

    Since Haswell will finally have DP 1.2 support and is very focused on power, these are the obvious items to address in the Redwood Ridge update. Thunderbolt controllers are pretty huge, and not necessarily designed on the same process as the chipsets they accompany, so integration may not be a reasonable option at this juncture.

    Thunderbolt has been shipping for over 2 years now, and brought 10 Gbit/s per lane serial I/O to consumer products while costing about an order of magnitude less than any other 10 Gbit/s technologies available. The cheap you're talking about doesn't happen for technologies that far out on the performance curve. If you want it now, you're gonna pay for it.
    Reply
  • watersb - Tuesday, April 09, 2013 - link

    excellent points. thanks, repoman27! Reply
  • dealcorn - Monday, April 08, 2013 - link

    It appears Intel's Avoton is the first chip scheduled for on SOC optical port support. Avoton is Xeon Atom @ 22 nm which is due prior to year end. If Intel migrates the technology to Atom desktop @ 14 nm about a year later, we may start to see more affordable options and perhaps an expanded role for diskless workstations with decent performance. Reply
  • mayankleoboy1 - Tuesday, April 09, 2013 - link

    External PCIE >> Thunderbolt. Reply
  • repoman27 - Tuesday, April 09, 2013 - link

    Yeah, in physical dimensions, I suppose. Go take a gander at the current crop of External PCIe cables and connectors and then have a look at a typical Ultrabook. Which interface makes more sense in that context? Reply
  • mayankleoboy1 - Wednesday, April 10, 2013 - link

    Which is cheaper to produce :) And which ises lesser power. And which is easier to fabricate. And which has all the bandwidth of the internal PCIE, not a subset. And which does not have any inter-between logic layers, so you get a pure PCIE interface, and leave the implementation to the specific device manufacturer. And which every OS since 2000 understands extremely well.

    TB is such a big fail that nobody is using it except for gimmicks.
    Reply
  • repoman27 - Wednesday, April 10, 2013 - link

    Idiotic troll is idiotic. Reply
  • mayankleoboy1 - Wednesday, April 10, 2013 - link

    Say that to the handful of TB supported storages and mobos. Reply
  • watersb - Tuesday, April 09, 2013 - link

    Now that USB 3.0 mostly catches up to FireWire for an external hard disk or two, ThunderBolt doesn't make sense to my parents.

    If ThunderBolt can last long enough to bring external GPUs to market -- the GPU embedded in the display, so that the ultrabook Just Works -- then people might want TB in addition to USB 3.
    Reply
  • geniusloci - Tuesday, April 09, 2013 - link

    USB3 doesn't support NCQ, TB does. Reply
  • daneren2005 - Wednesday, April 10, 2013 - link

    Quick, go tell grandma why she gives a damn about NCQ Reply
  • ATWindsor - Tuesday, April 09, 2013 - link

    Thunderbolt is interesting for me, but mainly becuase of range, imagine having a sentral PC and pushing video and touch/keyboard informaton over a (long) cable, one central HTPC for all displays, touchscreens in the wall for house-controll who communicate with a central server and so on. I feel intel is mainly to blame, one of the biggest pluses of the Tech is long range, optical thunderbolt didn't arrive before long after thunderbolt was released. Reply
  • Hrel - Tuesday, April 09, 2013 - link

    Simply cannot care about thunderbolt until it becomes ubiquitous. Why did floppy beat out zip drives? Cost and availability. Come on Intel, you gotta give up margins on this one or it will never happen. I dream of a world where ALL laptops have docking ports, docking ports so fast we can have external GPU's stored in our home docs. Docking ports so fast we can have home networking servers (in every home) hooked up to the same dock. A world where I can use my 15 year old dock on my brand new laptop. (granted, with less bandwidth). You'd think IBM would have made this abundantly clear by now, cross compatibility is how you BUILD a market. Don't try and be Apple, they will always be niche, be IBM. Reply
  • repoman27 - Tuesday, April 09, 2013 - link

    You kinda lost me with some of your analogies there. When did floppy beat out zip drives? In what market segments is IBM less niche than Apple at the moment? Would that be personal computers, operating systems, web browsers, tablets, smartphones, music or retail? Reply
  • Living Room Rocker - Friday, April 12, 2013 - link

    There seems to be a lot to be desired in understanding the usefulness of TB. At this moment I am using my TB port (on my G55VW laptop) to power and utilizes a(/n external) firewire audio card. (Try finding an affordable, Windows based laptop with firewire nowadays.) I am also anticipating a hub or docking station to connect two displays and multiple USB peripherals (including my firewire device) via TB to my laptop. All that using a single port on a laptop. It's just going to get better from here. Reply

Log in

Don't have an account? Sign up now