Will Thunderbolt upgrades ever exist for a 2012 Mac Pro?

[StaLevel]

This may have been covered before, and you are sure to let me know if that's the case, but it is an increasingly important question, and it would be nice to have a firm answer on it.

http://www.tomshardware.com/news/asus-thunderbolt-Z77-H77-pcie,15925.html

 

[ActionableMango]

No.

Even that Asus card only works on motherboards specifically designed ahead of time to work with it.

 

[Tesselator]

It's possible but I highly doubt we'll see anything of the sort this year. Maybe in 2015?

 

[deconstruct60]

Even that Asus card only works on motherboards specifically designed ahead of time to work with it.

Did that card ever pass Thunderbolt certification ? There were several "Rube Goldeberg" experiments/demos/etc that several motherboard vendors tried and showed in demos that never past certification.

Intel was clear from the beginning no PCI-e card solutions were on the horizon. From the live blog at Thunderbolts introduction

" ... 10:25 a.m. (Dong Ngo) : There won't be TB PCIe cards it seems. You'll need a new computer. ...."
http://news.cnet.com/8301-11386_3-20035571-76.html

There hasn't stopped folks from spending tons of effort into Rube Goldberg solutions of getting past the inherent constraint that need logic board based boot interface , PCI-e, and DisplayPort signals to do certified Thunderbolt host side implementation.

".... Inside the controller, you find a PCI switch and a collection of DMA engines collectively referred to as the Native Host Interface (NHI). ... "
http://www.tomshardware.com/reviews/thunderbolt-performance-z77a-gd80,3205-4.html

Cards in slots are not always going to be a viable solution for all problems.

----------

It's possible

Possible under the constraints of the passing the current certification standards... not really.

Possible for Intel to significant change to standard in the future, yes. Not very likely to happen. About as likely as hijacking the USB standard for a connector is. (not very much at all.)

 

[Tesselator]

Possible under the constraints of the passing the current certification standards... not really.

That should read: ...not easily.

 

[Macsonic]

@Deconstruct60 or anyone. Just have a clarification. Does this mean the 2013 Mac Pro may not have Thunderbolt ? That it all depends on Intel's revisions on cpu and motherboard but not likely to happen? Thanks.

 

[Tesselator]

@Deconstruct60 or anyone. Just have a clarification. Does this mean the 2013 Mac Pro may not have Thunderbolt ? That it all depends on Intel's revisions on cpu and motherboard but not likely to happen? Thanks.

I think we're talking about Thunderbolt on a card. Like this demonstration shows:

http://www.viddler.com/v/7af584b9

This is different than Thunderbolt on the system board.

 

[Macsonic]

Thanks Tesselator. My clarification was on the possibilty of the new 2013 Mac Pro natively supporting Thunderbolt. Not via card. Sorry if my post was not clear.

 

[Tesselator]

Thanks Tesselator. My clarification was on the possibilty of the new 2013 Mac Pro natively supporting Thunderbolt. Not via card. Sorry if my post was not clear.

No, I understood you clearly. The ney-saying here is about TB on a card I think. TB on a system board is already in existence. TB on a Xeon based system board is unlikely to appear in 2013.

 

[deconstruct60]

@Deconstruct60 or anyone. Just have a clarification. Does this mean the 2013 Mac Pro may not have Thunderbolt ?

It is possible. Thunderbolt doesn't really solve a major problem that a Mac Pro currently has. Mac Pro's have PCI-e expansion capabilities far in excess of even the future 2014 Thunderbolt bandwidth. They already have multiple video channel out capabilities to an actually a wider set of displays without having to pay for additional doodads. Mac Pros don't have major docking station issues to solve. Certainly nothing a integrated USB hub doesn't really largely solve.


It is far more Apple's obsessive compulsive disorder OCD tendencies that are driving the inclusion. It would make all Macs consistent (that doesn't pay attention to whether there is an issue or not with the system.) Because that is a Apple internal driven compulsion that makes the likelihood that it will be included pretty high.


Additionally, with an embedded GPU, Apple would be "free" to sell a Mac Pro with no GPU PCI-e card in it. It still would be a complete Mac system after plugging in a monitor ( so Apple's edict of only selling complete systems is happy) and there would be a bigger market for 3rd GPU cards. That would make the 3rd party card vendors and likely eventually users more happy. Apple gets to choose a GPU and users get more choose over what goes in the slots. Should make both sets of control freaks happy.
At least until they go OCD on what the "other guy" still has control over.

That it all depends on Intel's revisions on cpu and motherboard but not likely to happen? Thanks.

Intel doesn't make the motherboard for Mac Pro (well anymore. Initially they reportedly did).

The CPU is restrictions I don't really see where they are present. There is some spin that require a Intel CPU with Intel graphics. That makes it far more straightforward to implement since Display Port output is naturally on the logic board in that context. I haven't seen any evidence at all that it is a hard requirement.

No doubt that AMD is not touching Thunderbolt with a 10 foot pole, but I think that is largely because it is an Intel only part. Promoting it only puts more money into Intel's pockets. Primarily for the controller, but also the natural tie-in with integrated Intel graphics. Thunderbolt is going to expand Intel's graphics empire which is already bigger than AMD and Nividia. Those two aren't going to stop Thunderbolt getting traction in the market, but they aren't going to accelerate it either. I think their non-commital drives lots of handwaving about it being an Intel CPU/IO-Hub specific solution.

 

[ActionableMango]

Thanks Tesselator. My clarification was on the possibilty of the new 2013 Mac Pro natively supporting Thunderbolt. Not via card. Sorry if my post was not clear.

I assume the 2013 MP will have Thunderbolt, but that's not what OP asked when I said no.

 

[goMac]

No doubt that AMD is not touching Thunderbolt with a 10 foot pole, but I think that is largely because it is an Intel only part.

AMD is doing badly enough I think they'd do anything Apple asked them to.

Not that I think Apple has asked AMD to do anything.

 

[Macsonic]

Thanks to all for the clarifications. Appreciate the help.

 

[deconstruct60]

I think we're talking about Thunderbolt on a card. Like this demonstration shows:

Not really. That "card" is prototype backplane for a generic PC experimental container box. It has prototype TB controller that is being used for testing likey attached to a equally prototype classified reduced to a minimal subset logicboard. A 'jig' that someone is using to test some just initially fabbed silicon is not a finished product demonstration.


Normally you have to solder the TB controller to the logic board. If this "fresh from the fab" controller may not work you don't want to desolder and resolder the next one to the whole logic board to insert another prototype controller from next week's batch that comes back from the prototype fab. Probably want to take out the old "card", junk it and insert a new one. Otherwise evaluating a board subsubsection that has been tweaked multiple times. That is going to generation experimental noise over time.

That is not a normal PCI-e card. You can easily tell because it is not in a normal PCI-e slot. It is vertical instead of horizontal (with the box in this vertical orientation). There are few of the normal I/O sockets you would see on a standard backplane ( USB 3.0 , audio out, etc. ) present on that removable "card" backplane.

This is just as flawed as the folks making sweeping generalization from the prototypes seen in the Lightpeak demos. Ooh Lightpeak is going to come on a card just like this prototype. Not! Even the connector from the Lightpeak demos got dumped when it came to market. When seeing technology being demo 1-2 years ahead that is not standard PC parts and/or systems you are looking at.

 

[deconstruct60]

AMD is doing badly enough I think they'd do anything Apple asked them to.

If Apple asks AMD to shoot themselves in the head with Intel bullets they aren't going to. They aren't suicidal and frankly Apple probably wouldn't hand the that much component opportunity to make up for the overall market damage done.

AMD has worked on side shows such as :

http://www.anandtech.com/show/5413/amds-lightning-bolt-low-cost-thunderbolt-alternative-for-usb-30dp
( Obviously a bit before Apple launched "Lightning bolt" ) and pitching just straight USB 3.0 (inserted into core chipset before Intel ) .

AMD designs are generally the more affordable alternative. As long as TB controllers stay as relatively high priced selecting AMD and a more expensive controller just aren't going to show up in the same set of product designs.

Not that I think Apple has asked AMD to do anything.

In the x86 space. They should be asking them to do some things in the GPU space.

If AMD does have better integrated GPUs. They could ask AMD to some design bake-offs. If only keep Intel's 'only the paranoid survive' mode active. AMD has also had some good ideas they just had far more problems at execution and fabrication.

 

[Tesselator]

Not really.

There ya go with the strawmen again. You built your entire four paragraph response on the premiss that the existence of that card means or was implied as meaning, anything other than the fact it exists. Since no such implications were levied your response and personal attacks are completely irrelevant. Not unusual for you but I thought I would point it out this time for ya.

A bit like me shooting pics of the entirely wrong video card in that other thread - but regularly.

 

[deconstruct60]

There ya go with the strawmen again.

Always humorous how to you consistently label your technique as mine.

You built your entire four paragraph response on the premiss that the existence of that card means or was implied as meaning, anything other than the fact it exists.

So in answer to a direct question about Mac Pro 2013 TB solutions and Mac Pro 2013 updates to CPUs and motherboards you response of

" I think I think we're talking about ... "

has absolutely nothing to do with answering or responding to those questions at all. No connection what so ever.

I don't think that the vast majority of forum readers, especially new ones to the forum, are going to have that sort of expectation in a response to a direct question. That's why I wrote 4 paragraphs.

 

[Tesselator]

Always humorous how to you consistently label your technique as mine.

I'm rubber and you're glue? Really? Mmm-kay...

That's why I wrote 4 paragraphs.

Yeah, but all with wrong information.

 

[VirtualRain]

This may have been covered before, and you are sure to let me know if that's the case, but it is an increasingly important question, and it would be nice to have a firm answer on it.

http://www.tomshardware.com/news/asus-thunderbolt-Z77-H77-pcie,15925.html

Why is it increasingly important? TB on a Mac Pro seems like a solution in search of a problem. What would you use it for that can't already be done cheaper or faster (usually both) using the PCIe slots on a Mac Pro? Another way to look at is that all recent Mac Pros have dual TB "slots" built-in... No cables, adapters, docks, dongles, or enclosures required

 

[cathul]

Would be nice if i could use my thunderbolt display with a MacPro <sigh>.

 

[deconstruct60]

Would be nice if i could use my thunderbolt display with a MacPro <sigh>.

Your Mac Pro has a deep need for a docking station?

Primarily used as a monitor it is far less broadly compatible with alternative workstations and typically less capable on functionality ( color reproduction, input flexibility , etc. ) that many more focused display units on the market in the same price range.

The miniDP ACD is actually compatible with both Thunderbolt PC systems and the vast majority of the systems in the PC market. The TB display isn't.

Again this is one of those solutions in search of a problem. Sure there are folks with sunk costs into the display would will gravitate that way. Folks who want an Apple logo on everything they buy will be happier. But significantly large workstation market expanding through a huge value add proposition? Not really.

 

[Tesselator]

Very narrow thinking there.

The 10Gb/s would mean we could get nearly one gigabyte per second to/from our backup RAID boxes and TimeMachine volumes.

Who here doesn't want that - besides maybe someone who doesn't do anything with their Mac other than post nonsense on public forums?

And while SSD drives are currently topping out at around 600MB/s how long do you expect them to stay that slow?

And TB ports don't necessarily need to be external only.

S-T-R-A-W-M-A-N...

 

[cathul]

Your Mac Pro has a deep need for a docking station?

I know all that, but why should i need to buy a new display when i already own a thunderbolt display for more than a year? Why, at least, shouldn't a new MacPro don't have a thunderbolt connector so i can reuse my already existing thunderbolt display with it? Shall i sell the thunderbolt display just because i want a MacPro?

And having all those USB/Firewire ports on my desk is so comfortable that always reaching under my desk will soon be no option anymore.

 

[deconstruct60]

I know all that, but why should i need to buy a new display when i already own a thunderbolt display for more than a year?

If it is being used as a docking station now, why does that need disappear of buy a Mac Pro? Those current docking station duties can't possibly be for a current Mac Pro. That current device will loose its docking station functionality if dedicated the TB display to the Mac Pro.

I know the corner cases ..... " the MBP will be sold and replaced with Mac Pro". Or the "going to share a single TB display between laptop and Mac Pro because only use one Mac at a time" case. Those aren't mainstream use cases. If there was a solid need to do work in the field that drove the MBP purchase in the first place it is probably still there after get a new Mac Pro. Hence, still have MBP and still have need of a docking station. You already bought one, the TB Display. So going to buy another one?

Why, at least, shouldn't a new MacPro don't have a thunderbolt connector so i can reuse my already existing thunderbolt display with it?

The Mac Pro has a very small segment of the Mac market in the ballpark of 1%. If only 10% of that 1% needs fit these corner cases folks bring up that is not a market driver for inclusion on a Mac Pro. Especially for sunk cost equipment. The port should be adding something to create a far more expensive Mac Pro market. Not to primarily redundantly add the same kinds of expansion that the PCI-e already do.

Shall i sell the thunderbolt display just because i want a MacPro?

As noted above shall you sell what the TB display is attached to now because want a Mac Pro ?

If it is the wrong tool for the wrong job and don't have any use for it anything, then sure; why not?

And having all those USB/Firewire ports on my desk is so comfortable that always reaching under my desk will soon be no option anymore.

How is the TB display going to reach your Mac Pro under the desk. It has a permanently attached approximately 3ft long. For many setups it isn't going to reach. There are no male-to-female Thunderbolt extension cables. Not going to be any either ( not economical, nor worth decrementing the daisy chain count).

You can kludge the situations by buying a 2 or 3 meter cable but you are out another $50 and the permanent cable ( at least by these reports https://forums.macrumors.com/threads/1304576/ ) can't drive unpowered TB devices. All the more indicative that is primarily designed as laptop docking station.

 

[deconstruct60]

Very narrow thinking there.

The narrow thinking is that Thunderbolt alone solves these problems. The only "breadth" in thinking here is sweeping in MBA and Mac Mini solutions into the same solution space as the Mac Pro. For a discussion about what should go into Mac Pros that is a bit dubious if focused on maximizing the Mac Pro's potential.

The 10Gb/s would mean we could get nearly one gigabyte per second to/from our backup RAID boxes and TimeMachine volumes.

Anyone with a x4-x8 PCI-e SAS RAID card could achieve the ~1000MBps over 2014's 20Gb/s TB data transfer numbers in the recent Thunderbolt demo ( can't buy for a year or so ; just a demo) for the last 2-3 years. For example:

"...With its PCIe 2.0 interface, the MAXPower 4-Port eSATA RAID Card provides two 500MB/s data lanes for a total card bandwidth of 1000MB/ ..."
http://eshop.macsales.com/item/Newer Technology/MXPCIE6GRS4E/

or
" .... Without this, we simply couldn’t pull up numbers any where near 3GB/s write as seen here. ...."
http://thessdreview.com/our-reviews...ntroller-review-ssd-in-raid-0-as-ssd-hd-tach/


At some point Thunderbolt might compete against these solutions on price, but so far it has not to any clear and consistent advantage.

Who here doesn't want that - besides maybe someone who doesn't do anything with their Mac other than post nonsense on public forums?

Anyone who wanted it and had the money could already have it, instead pining away for Thunderbolt that isn't deployed yet. Thunderbolt in and of itself doesn't make for fast SSD RAID 0 arrays. There are multiple ways of achieving that. There is a difference between thinking broadly to find a focused solution to a problem and thinking broadly to grasp at justifications for Thunderbolt.

And while SSD drives are currently topping out at around 600MB/s how long do you expect them to stay that slow?

No time at all because not really restricted now for Mac Pro users.

http://www.sonnettech.com/product/tempossd.html

Although a internal only solution it is easily mutable into a external one.

And TB ports don't necessarily need to be external only.

This is one of those "sales pitch spins" from back in the LightPeak era. Now that TB is out, there little solid evidence to back that up.

" ... The spec for max trace length between the Thunderbolt controller and port is two inches, compared to up to 10 inches for Intel's USB 3.0 controller. ... "
http://www.anandtech.com/show/5884/...s-part-2-intels-dz77rek75-asus-p8z77v-premium

PCI-e and others have longer trace length restrictions than "raw" TB signals do. Sure, once you get to an amplifying transceiver the distances get bigger, but costs and internal space usage have jumped.

Back when Lightpeak had the possibility of building the transceiver amps into the controller that sales pitch made more sense. With the current design mandates of having active amp/transceivers in both side of the cable, it really doesn't make much sense in the PC space for internal deployment.

That kind of set-up requirements may make sense in a "large as a rack" computer, but not in the PC space. There are no 1-3 m distances to cover inside PCs in general. In fact, the general trend for PCs is to get smaller; not larger. So Thunderbolt doesn't really bring anything significant to the party distance wise. Fiber TB might cut down on internal RF problems. However, the fiber solutions for TB aren't all that much cheaper (relative to PC pricing norms) than other solutions already leveraged in larger computers already (bigger IBM boxes). [ Fiber being another one of these "Lightpeak" spins that hasn't show up in the marketplace. ]

S-T-R-A-W-M-A-N...

It is your wobbly weak proposals that are strawman.