They stopped selling it in 2012, so 5 or 6 years is 2017 or 2018.
The 2008 is still supported by updates, and that's at the 7 year mark. If they support it again this year, it'll be 8 years.
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.
2016 nMP
- Thread starter ManuelGomes
- Start date
-
- Tags
- nmp
- Sort by reaction score
- Status
They stopped selling it in 2012, so 5 or 6 years is 2017 or 2018.
The 2008 is still supported by updates, and that's at the 7 year mark. If they support it again this year, it'll be 8 years.
It seems with the current fixation on leveraging the GPU as much as possible in Pro apps (and Metal's release), and the stagnation of CPU progress, that the main reason to wait to upgrade the Mac Pro had to do with AMD's progress with their GPU tech. Broadwell Xeon's are coming out, and North Bridge controllers aren't dependent on having Skylake architecture, so I don't see CPUs as being a reason to hold back since they'll be able to use TB3. The GPUs meanwhile are where the real progress can be found.
What do you mean TB3 is free with Skylake? Nothing is free, and TB3 is supported on Broadwell as well or which ever platform you can squeeze x2 or x4 PCIe 3, depending on the SKU 8or number of ports) you want.
Skylake has the advantage of having more PCIe 3 lanes available (on the PCH), that's all. It's not a requirement for TB3.
I started reading the article but man, do they even know what they're talking about? It seems they just slapped some bits and pieces they got from somewhere else and made an article. What a mess.
Skylake has the advantage of having more PCIe 3 lanes available (on the PCH), that's all. It's not a requirement for TB3.
I started reading the article but man, do they even know what they're talking about? It seems they just slapped some bits and pieces they got from somewhere else and made an article. What a mess.
No they didn't stop selling them in 2012 I have a 2014 5,1 with AC good until 06/11/17.
I may have one of the last ones but they were still selling in 2014.
Indeed, CPUs are not the issue here. GPUs are, and I believe they're sorting out, or even considering if they should do it, this TB3 and DP1.3 problem. They probably really want to release something that is not a workaround but rather a proper solution, no compromises. And officially this is not possible, at least that I know of.
Intel supports only DP1.2 in TB3, so either they do a TB3.1 update or Apple does something proprietary.
The reasonable solution, although not the best, is doing 2 DP1.2 MST, as on the iMac. But they know this is not optimal, and I'd also like to see a SST solution.
That's why the nMP keeps getting delayed, no tech available yet for the machine they want to build, perfect as they see it.
Intel supports only DP1.2 in TB3, so either they do a TB3.1 update or Apple does something proprietary.
The reasonable solution, although not the best, is doing 2 DP1.2 MST, as on the iMac. But they know this is not optimal, and I'd also like to see a SST solution.
That's why the nMP keeps getting delayed, no tech available yet for the machine they want to build, perfect as they see it.
http://gpuopen.com/welcometogpuopen/
Good lecture.
Interactive library for compute tools, libraries, etc, etc.
Well played AMD, well played.
Good lecture.
Interactive library for compute tools, libraries, etc, etc.
Well played AMD, well played.
Last edited:
Now if only they would spend a little more time on their drivers everything would be peachy
There is an misconception, TB3 doesn't means Dp 1.3 neither usb-c means Dp 1.3. (dp alt mode).
In fact Dp 1.3 only can be implemented over usb-c on cables upto 1mt (3ft) long, which makes it practically unusable. READ
http://www.anandtech.com/show/8558/displayport-alternate-mode-for-usb-typec-announced
Tb3 have never intended to implement dp 1.3 and neither Tb4 will do unless Intel move Thunderbolt to another connector.
The big difference with previous Tb2 is that you can have 2 Dp 1.2 signals on the same cable allowing 5K@60hz over a single cable. (right now you can drive a 5K display from a Mac Pro using 2 dp cables simultaneously, exactly the same setup can be done on a single tb3 cable).
There are also misconceptions about MST and SST, and while earlier implementation of MST may penalize either the GPU and the display, this is matter from the past on newer implementations (specially if they support 5K from the beginning not from driver update, which means the hw is optimized to deliver MST signaling).
Only bottlenecks maybe the HDCP which may not work for 5K on MST.
So I don't think if ever usb-c will support Dp 1.3, and indeed the near future will be difficult to find GPU supporting dp 1.3 unless using another interface.
Last edited:
Funniest part is that on other forums people already asked: "How long till people will say that AMD should now work on their drivers?"Now if only they would spend a little more time on their drivers everything would be peachy
Funniest part is that on other forums people already asked: "How long till people will say that AMD should now work on their drivers?"
As I said in another thread, I'm mostly agnostic when it comes to OS and I've both used AMD and NVidia interchangebly over the year. But the drivers situation in regard to AMD on Linux is too painful to endure. I know that they have started a new project concerning those but they don't assign enough ressources to it in my opinion making NVidia the only real choice for hardware accelerated graphics on Linux.
AMD drivers in Linux are a$$tastic Windows drivers though are pretty good, especailly their FirePro drivers
AMD released full new drivers full open-source AMDGPU, the only catch is that those drivers are only for newer models, (Fiji, Carrizo, Tonga, Stoney, and Iceland chipset).
There are open source and there are proprietary AMD drivers... Both are subpar.
[doublepost=1453838701][/doublepost]
Are you basing your opinion on a single use case, yours?
Do a search or go to some Linux forum and ask around you'll see what I mean. Eh, I'm not here to bash AMD on this, I'm only acknowledging a plain fact, that on Linux AMD drivers suck. Maybe they could fix them by putting a few more dev to debug them...
You're making a misstatement here. USB-C of course supports DP 1.3. All usages of USB-C whether for DP or TB3 have issues related to length of passive cables. Longer runs will require active cables.
It is not only my case. Currently there is a stack of complaints about Nvidia GPU display drivers crashing constantly regardless of platform. Today I have watched a gameplay footage captured with dual GTX980's when the problem appeared while it was captured. I read today a thread about GTX970 loosing display signal thanks to faulty drivers.
All I said is that BOTH of GPU producers have to work on their drivers. I did not disagreed with you on the topic of AMD drivers, because I do not have knowledge of that. But I do have a lot of knowledge what is going on with Nvidia drivers currently.
Again those are single point on a chart, not a proof of instability. Most of those gamer are overclocking their cards passed what is recommended. This is what brings instability, not the drivers.
But AMD lack of ressources mean that while the GPU does work, the acceleration is piss poor meaning that you can't use them for anything beyond displaying the desktop. Forget about games or graphic intensive application. They work, just like a old Yugo car worked... Very slowly...
I will completely ignore the part about AMD, because it is not my part of interest. I do not simply have knowledge about it.
No. The problem with loosing display signal is caused by faulty Nvidia drivers. The problems with display drivers crashing is caused by faulty drivers. Threads on Nvidia forums are full of pissed people who for example RMA'd GPUs for 4 times in a row and still got the same thing. The clocks have not been touched. I have not OC'ed my GPU and still got driver crashes. Is it also my fault that my drivers are crashing constantly?
I didn't say TB3 is DP1.3 or anything like that. DP1.3 is supported in DP Alt Mode over USB-C, but not officially so in TB3, this is what I was saying.
Intel or Apple could make it work also on TB3, I guess, if they wanted to.
I could see a TB3D with a 1m passive cable, or longer active one. Would you see a problem here?
Intel or Apple could make it work also on TB3, I guess, if they wanted to.
I could see a TB3D with a 1m passive cable, or longer active one. Would you see a problem here?
All I can say is that those are point on a graph... I have presently about 20k PCs running with NVidia GPU where I work with zero drivers problems... Many thing can crash a drivers. An ill behave application that does an illegal call or access a wrong part of memory can crash a drivers. Those problem aren't with the driver but with the OS and/or application. And again, I'm not saying that AMD drivers are unstable, I'm saying that they perform like an earthworm going through molasses in cold temperature...
Please read the link from arstecnica "DP alt mode is only supported on passive cables"
USB-C as Thunderbolt 3 are intended mostly for consumer applications, and 5K Display are associated with pros, notwithstanding the challenge a single usb-c can drive a 5K display just only on MST and disabling all other functionality (usb or Thunderbolt), that will be the price to support 5K on USB-C, That's why I insist the next 5K cinema display will not be Thunderbolt (unless has an integrated gpu to offload the host from GPU processing).
Whatever the nnMP will have 4 TB3/USB-C and another 6 USB3 (maybe having 2 usb-c non-thunderbolt but with DP alt mode capabilities to enable 6 5K displays as the nMP can drive up to 6 4K).
So, DP on USB-C/TB3 is limited to passive cables either DP 1.2 or 1.3, the latter only can run 3FT on usb-c cables which makes it impractical unless you wanna dock an laptop or an pc at this short distance.
USB-C as Thunderbolt 3 are intended mostly for consumer applications, and 5K Display are associated with pros, notwithstanding the challenge a single usb-c can drive a 5K display just only on MST and disabling all other functionality (usb or Thunderbolt), that will be the price to support 5K on USB-C, That's why I insist the next 5K cinema display will not be Thunderbolt (unless has an integrated gpu to offload the host from GPU processing).
Whatever the nnMP will have 4 TB3/USB-C and another 6 USB3 (maybe having 2 usb-c non-thunderbolt but with DP alt mode capabilities to enable 6 5K displays as the nMP can drive up to 6 4K).
Further, forget all that discussion on DISPLAYPORT, HDMI 2.0 supports 5K SST at 60hz also with HDCP.
So I fear my friends you should save your frustration and accept the nMP as other Thunderbolt 3 enabled device will not come with DP 1.3 UNLESS Intel revise the standard and enable it for 1meter (3ft) long USB-C cables.
Next AMD GPU will support both DP 1.3 and HDMI 2.0 but maybe not the case of Thunderbolt 3 unless Intel's give up the 1 Mt restrictions.
So I fear my friends you should save your frustration and accept the nMP as other Thunderbolt 3 enabled device will not come with DP 1.3 UNLESS Intel revise the standard and enable it for 1meter (3ft) long USB-C cables.
Next AMD GPU will support both DP 1.3 and HDMI 2.0 but maybe not the case of Thunderbolt 3 unless Intel's give up the 1 Mt restrictions.
You can't run a 5K display over a longer passive TB3 cable than you can a DP1.3 cable, full stop. The Intel documentation supports this.
- Status