Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

New Mac Pro Announced at WWDC! - It's OFFICIAL!

Maybe I'm not understanding something with the Thunderbolt 2 ports. There are six ports, but I feel like I distinctly remember them saying in the keynote that they're backed by three Thunderbolt 2 chipsets.
Each Thunderbolt 2 host chipset supports two ports.

Wouldn't that mean 12 lanes are backing the ports, not 24? Meaning you can't have all six ports simultaneously receiving full TB2 bandwidth?

How they are dividing up PCIe lanes is still uncertain. There is also some talk that some devices in the system, possibly the GPUs, are not hosted via primary PCIe channels, freeing up those lanes.
 
Each Thunderbolt 2 host chipset supports two ports.
Yes, I guess my question was more along the lines of: would two devices, plugged into each of the ports on one chipset, both get full TB2 bandwidth? I imagine connecting one to a monitor, and one to a device(s) wouldn't cause a problem, because of how the channels work - but would, say, a GPU connected to each port be any different than connecting both to one port?
 
Yes, I guess my question was more along the lines of: would two devices, plugged into each of the ports on one chipset, both get full TB2 bandwidth? I imagine connecting one to a monitor, and one to a device(s) wouldn't cause a problem, because of how the channels work - but would, say, a GPU connected to each port be any different than connecting both to one port?

The controller can supply full bandwidth to both ports simultaneously, provided they are connected to a big enough PCIe pipeline on the back end -- 16 lanes of PCIe v2 (8 lanes per port) or half as many lanes for PCIe v3. So now we wait for Apple to unveil all the specs and see what's up. As I mentioned in the above post, there are some rumors flying around, but many of them conflict. It's entirely possible that they could have rolled the Thunderbolt controllers into a new addition to the C600 series chipset family, incorporating TB2 hosts directly onto the CPU data bridge, rather than allocating individualized lanes on the CPU's PCIe bus. Of course, the CPUs only have a certain amount of bandwidth, but there are alternative ways to allocate rather than integrating within the standard confines provided by the CPU's peripheral connections. This is a different design approach, we don't know yet what we have to work with or how lanes and bandwidth are being allocated. It's also possible that Apple and Intel could be using customized versions of the E5 V2 CPUs to give the connectivity they need. I would be rather surprised if this is the case, but it's happened in the past with the Macbook Air and customized CPU and chipset combos. Apple could sell more than enough of these boxes to make it feasible.

As of this point, it's all academic speculation. What we do know is that they have demoed working Mac Pro systems running 12-core Ivy Bridge EP V2 Xeon processors. The systems have been touched by Pixar, Luxology/ The Foundry, Autodesk, BlackMagic and Adobe and all those companies are giving a thumbs-up. So they can't be all bad...
 
It just seems like exactly what I want for onset data management and something less than what I need for demanding post-production work. Can't wait to see a real one in action.
 
The difference between Thunderbolt 1 and Thunderbolt 2 is that v2 has 4 x 10Gbps channels instead of just two. We will need new cables and new devices to take advantage of it. It's not 100% clear, but it seems that when connecting TB1 devices (including displays) that TB2 ports go into 10Gbps per channel mode. I don't know how that plays into TB2 device use and it may be wise to keep TB2 devices on certain ports and keep all TB1 devices on different ports. That said, when operating in TB2 mode, one of the 20Gbps channels is dedicated to the video stream, uni-directional. All other data must pass over the other 20Gbps channel, same as in TB1, just double the bandwidth. In TB1, you connect a display and your TB port now only has one 10Gbps channel to work with.

Again, where are you getting your information? This seems to be at odds with what Anand says over at Anandtech, although I could simply be misunderstanding you.

TB 2 represents NO increased in overall bandwidth.

From the article:
"Today, Thunderbolt exists as four 10Gbps channels - two upstream and two downstream. Each channel however is fully independent. Although PCIe and DisplayPort are muxed from the cable perspective, you can only send one or the other over each channel. That limits max performance for a single storage device to 10Gbps (minus overhead), and it similarly limits the max display bandwidth to 10Gbps as well...
By combining the channels together, Thunderbolt 2 enables two 20Gbps bi-direction channels instead of two sets of 10Gbps channels. There's no overall increase in bandwidth, but the solution is now more capable."

http://www.anandtech.com/show/7049/intel-thunderbolt-2-everything-you-need-to-know

Those 2 separate 20 Gb/s channels Anand is referring to are an upstream and a downstream channel, not two upstream AND downstream channels. Hence there is only 20 Gb/s bandwidth (much less in reality when you count the PCIe overhead) for a given purpose. Which also means that each controller (which powers 2 x TB 2 ports) will need to be able to handle 40 Gb/s. This is fed by a x4 link to each TB controller. (See here:https://twitter.com/anandshimpi/status/348582181960044545) Which means that the graphics cards are likely in a x8 configuration.
 
Anand / Anandtech doesn't have all the right info, that's already been discussed here and in other locations on the web. I have made some errors too as I was going off of what I was being fed from others who were being spoon-fed info.

I can't go back and elaborate and correct now other than to say I made a couple goofs. Not all of the correct info is out there. I'm now part of a group working on a new Thunderbolt product and I do have more info. I also have a non-disclosure commitment, so that is that. It will all come out soon enough. The Apple guys made a couple goofs in their presentation as well. Kinda funny, or perhaps even by design?

As for the new Mac Pro, word on the street is Apple has pulled the demo units and no one outside the mothership is working on them at the moment. Not sure what that means, if anything. I think Apple just put a few in the wild with high-profile companies they can trust and invited a few key companies to preview them just before WWDC -- Pixar, Autodesk, TheFoundry, etc.. All to get the word out there. Beyond that, I think everything goes back under wraps until they ship.
 
As for the new Mac Pro, word on the street is Apple has pulled the demo units and no one outside the mothership is working on them at the moment. Not sure what that means, if anything. I think Apple just put a few in the wild with high-profile companies they can trust and invited a few key companies to preview them just before WWDC -- Pixar, Autodesk, TheFoundry, etc.. All to get the word out there. Beyond that, I think everything goes back under wraps until they ship.

Considering how notorious Apple is for making it so even their own engineers sometimes don't know exactly what product they are working on, I wouldn't read too much into Apple pulling back the demo units. Even having demo units outside the mothership at all before release date is very un-Apple in relation to how they operated under the second reign of Jobs.

So I sit here in Wait-And-See mode. Dunno if I'll jump on the next MBP to run my DITing off, or hold out for a new Mac Pro and build out a large Pelican case with a screen built into the top, thunderbolt raid, and Mac Pro that just has to be stood up and turned on. The latter will probably cost me about twice as much, but if I get the work to warrant it, then it's money well spent. I really don't want to build up a full cart, because I have no way to transport a cart. But a large pelican can fit in my trunk, and setup wherever. I think the tech is there now that for most jobs below the big big features, a cart isn't necessary. Look at the LilyPad Case from LightIron. I'd choose that over a current Mac Pro (one that doesn't have the same software, of course) for absolutely any job.
 
At this point, as Jeff notes, there is enough conflicting info out there about the internals of the new MacPros that breaking down specs past a certain point is not particularly illuminating. What we do know, is that every system has it's limitations and even the most optimistic guesses about the new cylinder would not equal a Pablo. FWIW, I think the real world responsiveness of the new design will be excellent due to more direct connectivity to the mobo and the uber fast flash storage.

IAC, what I am very curious about is what kind of options we will have to use several of the little cylinders together in a render farm config. Would the TB2 I/O bandwidth support distributed rendering well enough to make it feasible? If you put 6 of the new cylinders on a custom rack shelf could you create a cluster capable of Clipster/Pablo level performance?

FWIW, the other element of the equation for working with high resolution media files is the format of those files. 4K OpenEXR 32bit half float image sequences present a different set of requirements for the hardware than compressed formats like Cineform, ProRes 444 4K, native R3Ds, H.xxx, etc. Clipsters, Pablos and other high end shredders have typically been optimized for 10bit log DPX media and/or 16bit linear TIFFs. Will those formats still be the most relevant going forward?

What I'm hoping is that a single Mac Pro cylinder will be plenty of power for 90% of my needs and able to be clustered easily when more muscle is required.

Cheers - #19
 
Anand / Anandtech doesn't have all the right info, that's already been discussed here and in other locations on the web. I have made some errors too as I was going off of what I was being fed from others who were being spoon-fed info.

I can't go back and elaborate and correct now other than to say I made a couple goofs. Not all of the correct info is out there. I'm now part of a group working on a new Thunderbolt product and I do have more info. I also have a non-disclosure commitment, so that is that. It will all come out soon enough. The Apple guys made a couple goofs in their presentation as well. Kinda funny, or perhaps even by design?

As for the new Mac Pro, word on the street is Apple has pulled the demo units and no one outside the mothership is working on them at the moment. Not sure what that means, if anything. I think Apple just put a few in the wild with high-profile companies they can trust and invited a few key companies to preview them just before WWDC -- Pixar, Autodesk, TheFoundry, etc.. All to get the word out there. Beyond that, I think everything goes back under wraps until they ship.

Gotcha. Thanks for the (somewhat clarifying) clarification.

Is there a readily accessible place that points out the specific errors?
 
AHHH fuck me, just typed a long researched post...

Basically some people can get privy to a lot of info (like Jeff), or me when I was a VAR and get a little ahead
of ones self when talking on tech. It's not on purpose, and NDAs can be a bitch, but it's easy to talk about
future tech you may have seen or a buddy etc., will be coming out, but never makes it out of the LAB.

Happened more than once to me back in the day. Also as a VAR got caught between clients and manufactures
cause the f***ing tech didn't work cause some other manufacture in the supply chain didn't exactly build their
components to full spec, negating what company X built pushing the boundary of the spec, but failing when
including under spec technology.

Apple could tweak the spec, but backwards compatibility would be trashed if you went with 40Gbs bidirectional (80Gbs aggregate), the cables now I doubt could handle signaling that high.

So maybe they will allow grouping of ports to function as one, or maybe they are just putting out marketing that on it's face "while technically right" is really a bit misleading.


Anyways, had a ton of techy links beyond just Toms, Anand, and Ars but lost them when I F*ck up and pushed the wrong button.

With a little searching you can find those articles from more drab tech sites that have "IEEE Fellows" and steering committee members who talk of possible expansion and signalling increase above the initial thunderbolt 2 spec.

Start here (you might already been there), I am too tired to do it all over again and post all the other links.

http://ark.intel.com/compare/71880,71879,66000,67022,65998,66003,65995

or
http://blogs.intel.com/technology/2...ndwidth-enabling-4k-video-transfer-display-2/


http://www.intel.com/content/www/us/en/io/thunderbolt/thunderbolt-technology-developer.html
 
Another good reason why hardware should be opened the new quadro k6000 just announced.
 
New MACPRO and Bigfoot Mobile Systems rackmount options and carts

New MACPRO and Bigfoot Mobile Systems rackmount options and carts

Hey guys,
Not sure if this thread is still warm but wanted to get the word out that we have designed an elegant and funtional “Rackable” system option for the new MACPRO’s and Thunderbolt goodies. This option will take the up same 5.5-6RU space as the current MAC laying on its side, so it will take up the same space or a little less width as the old MAC. There will be a couple of layout options in either one of our aluminum drawer slide outs or a fixed aluminum rack-mounted shelf.
Option 1 will include up to 2each MACPRO’s standing up along with desktop 2-4 slot Expander and desktop RAID all in the 5.5- 6RU’s rackspace. Our depth is 22” or 24” so plenty of real estate. Vented base will allow for better than a desktop airflow under the MAC which is essential on this design. The other non-rackmount Thunderbolt appliances will be held in as well as the MACPRO’s with “no-hardware” needed and will provide a secure and shock mounted system in the design. Option 2 would be 4 each MAC’s, with Option 3 allowing for up to 6 each MAC’s, all in the same space allocation. There will also be an option to lay the MAC’s down as well. Two per drawer if height needs to be lowered to 4RU’s.

Any of our current clients will be able to upgrade or replace their original MACPRO drawer as well as all new carts.

Along with this we are excited to introduce some very recent cart designs which I will post soon that have shipping already. These are extremely durable and functional as current Bigfoot carts are known for. The new designs are extremely flexible (modular split-apart cabinets and base), lighter, trimmer, and also include new wheel/base options. These and our roll-able/carry-able mini racks (Versa Flypak, single or dual rack) clamshell style cabinets and cases will also be posted soon.
Please email or call direct or post here if you have you need any more information or have questions etc.

All the best!
Doug Solis
(707) 333-4400 Direct

President
Bigfoot Mobile Systems/
Mobilized Tech Systems

C:\Users\DOUGSO~1\AppData\Local\Temp\msohtmlclip1\01\clip_image002.jpg


www.bigfootmobilecarts.com
doug@bigfootmobilecarts.com
(888) 293-0869 Toll Free
 
Doug,

Do you have any pictures to share?

Hey guys,
Not sure if this thread is still warm but wanted to get the word out that we have designed an elegant and funtional “Rackable” system option for the new MACPRO’s and Thunderbolt goodies. This option will take the up same 5.5-6RU space as the current MAC laying on its side, so it will take up the same space or a little less width as the old MAC. There will be a couple of layout options in either one of our aluminum drawer slide outs or a fixed aluminum rack-mounted shelf.
Option 1 will include up to 2each MACPRO’s standing up along with desktop 2-4 slot Expander and desktop RAID all in the 5.5- 6RU’s rackspace. Our depth is 22” or 24” so plenty of real estate. Vented base will allow for better than a desktop airflow under the MAC which is essential on this design. The other non-rackmount Thunderbolt appliances will be held in as well as the MACPRO’s with “no-hardware” needed and will provide a secure and shock mounted system in the design. Option 2 would be 4 each MAC’s, with Option 3 allowing for up to 6 each MAC’s, all in the same space allocation. There will also be an option to lay the MAC’s down as well. Two per drawer if height needs to be lowered to 4RU’s.

Any of our current clients will be able to upgrade or replace their original MACPRO drawer as well as all new carts.

Along with this we are excited to introduce some very recent cart designs which I will post soon that have shipping already. These are extremely durable and functional as current Bigfoot carts are known for. The new designs are extremely flexible (modular split-apart cabinets and base), lighter, trimmer, and also include new wheel/base options. These and our roll-able/carry-able mini racks (Versa Flypak, single or dual rack) clamshell style cabinets and cases will also be posted soon.
Please email or call direct or post here if you have you need any more information or have questions etc.

All the best!
Doug Solis
(707) 333-4400 Direct

President
Bigfoot Mobile Systems/
Mobilized Tech Systems

C:\Users\DOUGSO~1\AppData\Local\Temp\msohtmlclip1\01\clip_image002.jpg


www.bigfootmobilecarts.com
doug@bigfootmobilecarts.com
(888) 293-0869 Toll Free
 
Back
Top