Does anyone know yet if Thunderbolt 2 allows direct mini dp connection?

[Zellio]

Instead of the mini dp at the end of the chain?

 

[Mikey7c8]

Pretty sure the limitation was on the display, not the tb port

 

[zaphoyd]

Based on my knowledge of the way Thunderbolt displays work Thunderbolt 2 will not change the mechanics of how DisplayPort is passed through except that the passed through DisplayPort signal will be a DP1.2 signal rather than 1.1.

I would expect:
- Mini DisplayPort monitor will still not be able to be connected directly to a Thunderbolt 2 monitor that is plugged into a Thunderbolt 2 port.
- Mini DisplayPort monitor will continue to be able to be connected to the end of a Thunderbolt 2 chain where the last device is not a monitor (or some other device that uses a DisplayPort signal like some docks)
- A 60Hz 4K display (and hopefully also an MST chain of smaller DisplayPort displays) will be able to be connected anywhere that Thunderbolt 1 would allow a DisplayPort monitor.

 

[deconstruct60]

Instead of the mini dp at the end of the chain?

If "direct connection" means plugging the mDP (mini DisplayPort ) device directly into a computer system's Thunderbolt (TB) ports, then yes. But it is not "yes" in the notion of 'instead of'. If there are no other TB devices, that is the end of the chain. A chain length of zero ( or 1 if count the starting device), but it definitely the end. Neither TB 1 or 2 have any problem with that.

Based on my knowledge of the way Thunderbolt displays work Thunderbolt 2 will not change the mechanics of how DisplayPort is passed through except that the passed through DisplayPort signal will be a DP1.2 signal rather than 1.1.

But DP1.2 can have multiple streams in it. So that is possible difference.

I would expect:
- Mini DisplayPort monitor will still not be able to be connected directly to a Thunderbolt 2 monitor that is plugged into a Thunderbolt 2 port.

The 'problem' with a TB display is that the DP decoder and coupled DP handling subsystem could only hand one display stream. Since the device itself needs to consume one and only one comes out of the TB controller there isn't another decoded stream to pass to a second TB port (the one that wasn't the input downstream from the computer).

In a DP1.2 context there could be two display streams inside the TB monitor device. Since it only needs one the second couple possibly be fed back into the TB controller for output on the second port. It depends upon if Intel wanted to add that additional complexity into the controller and also whether impose that additional complexity onto the non-display TB devices that need to funnel DP1.1/1.2 to devices on their 'other'/2nd port.

- Mini DisplayPort monitor will continue to be able to be connected to the end of a Thunderbolt 2 chain where the last device is not a monitor (or some other device that uses a DisplayPort signal like some docks)

As pointed out initially above, the "last" device can be the first device; chain length zero is still the end.

- A 60Hz 4K display (and hopefully also an MST chain of smaller DisplayPort displays) will be able to be connected anywhere that Thunderbolt 1 would allow a DisplayPort monitor.

That is even less limiting of TB chain bandwidth if just direct connect to the computer sytsem's TB ports.

However, a 60Hz 4K display probably would choke off any other displays on a TBv2 chain. It will be a similar effect as no display output of a TB display device, only it would also impact everything else down the chain.

 

[zaphoyd]

If "direct connection" means plugging the mDP (mini DisplayPort ) device directly into a computer system's Thunderbolt (TB) ports, then yes. But it is not "yes" in the notion of 'instead of'. If there are no other TB devices, that is the end of the chain. A chain length of zero ( or 1 if count the starting device), but it definitely the end. Neither TB 1 or 2 have any problem with that.

Sure. Directly plugging into the computer TB port puts it in DisplayPort passthrough mode where there isn't really a Thunderbolt chain at all. The main difference here with TB2 is that DisplayPort 1.2 passthrough is now supported. In theory the 2013 TB1 controllers should allow this as well, though Apple seems to have not used them in any of their 2013 macs so far. =(

The 'problem' with a TB display is that the DP decoder and coupled DP handling subsystem could only hand one display stream. Since the device itself needs to consume one and only one comes out of the TB controller there isn't another decoded stream to pass to a second TB port (the one that wasn't the input downstream from the computer).

In a DP1.2 context there could be two display streams inside the TB monitor device. Since it only needs one the second couple possibly be fed back into the TB controller for output on the second port. It depends upon if Intel wanted to add that additional complexity into the controller and also whether impose that additional complexity onto the non-display TB devices that need to funnel DP1.1/1.2 to devices on their 'other'/2nd port.

I hadn't thought of the ability to use DisplayPort MST on a Thunderbolt display to feed the signal back into the TB chain. Somehow I doubt anyone will actually do this.. Apple has no reason to want to allow you to buy anything other than more Apple Thunderbolt displays which wouldn't need such a feature. If they do though it would be much appreciated. If anyone other than Apple would make a TB display that would be awesome also.

However, a 60Hz 4K display probably would choke off any other displays on a TBv2 chain. It will be a similar effect as no display output of a TB display device, only it would also impact everything else down the chain.

Yes, it is likely that a 60Hz 4k display at the end of a Thunderbolt chain would result in severely limited outgoing (from the computer) bandwidth for the rest of the Thunderbolt chain. I wonder how much the bandwidth limitations will affect any 4k Thunderbolt monitor that Apple might make. There wouldn't be much bandwidth left for the webcam/speakers/usb/firewire/ethernet ports on the existing ones much less for any devices in the chain after the display.

I also wonder how TB2 flow control/QoS will work with respect to displays and audio devices now that the two channels are combined. Will a TB monitor on a TB2 chain with a fast SSD array have video or audio issues when the SSDs are being accessed at near TB2 line speeds? Does position in the chain imply priority? Something else?