Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
It only means a Bunch of rMBP and iMac retina, are coming very soon, same dGPU on current retina lineup should correspond to each of these, the good news: an earlier updated iMac Retina 5K is coming, bad news still no Mac Pro update, at least with beta OS support, maybe Apple is using web drivers for the nnMP prototypes, also this only mean no nnMP release until at least macOS 10.12.02 (or about Q4'16 Q1'17).



This info is confusing and illogical, need to re-check maybe Apple is introducing a camouflaged GP100 for the Mac Pro, Apple has no need to support OLD nVidia HW not sold on they machines, which Mac will host this "old" nVidia GPU ?



That's disappointing but no surprises here. Hopes AMD/Intel/Khronos.org launches its own OpenCL 2.x support files and toolchain for OS/X.



Metal-Metal-Metal.... I do prefer classic rock...
You know, mago. What if people were digging something that doesn't exist? Just saying.
Maybe Apple will announce it or maybe won't. But at least there's some hope for 6,1.
 
Not all blind people are totally blind.
Many have different degrees of sight impairment
that the bright, high rez and immersive optics of
VR will allow them to visually experience life in a way
that they never have before.

Take off your blinders and think different.

When Apple finally, if ever, allows the obedient herd to have VR, the naysaying faithful
will suddenly about face and woo woo that it's the greatest thing since sliced bread.
Exactly.
 
...

---
OpenGL/OpenCL have not been updated. Still 4.1 and 1.2, no new extensions or anything.
---

...

Source:
netkas.org
[doublepost=1466265176][/doublepost]It seems that Metal works now on every Mac that support macOS Sierra.. and OS' UI runs only on Metal, openGL is not supported anymore... but I have to confirm this.

Technically there is a difference between no feature updates and not supported. I'm pretty sure I saw on a WWDC 16 slide that Apple has layered Metal under OpenGL ES on iOS. (either in "What's new" https://developer.apple.com/videos/play/wwdc2016/604/ or in "Adopting" https://developer.apple.com/videos/play/wwdc2016/602/ ) Metal is low leveled enough there should not be a big problem in putting mainstream OpenGL on top for macOS over a relatively long term. Especially the way Apple splits OpenGL implementation into the Apple "top half" and the GPU vendor "bottom half". The GPU vendors were already getting the "bottom half" anyway.

Metal is more parallel ( not one big state ) , more context ( single context ) , lower level , less error checking. There are no conflicts from the feature table here:

https://en.wikipedia.org/wiki/Vulkan_(API)#Features

with OpenGL on left always being the more "higher level" of the two.


There is some translation of shader code Apple might have to do to get a broader scope of OpenGL, but those tended to be the parts Apple has historically be hesitant to enable anyway (i.e., the "no new extensions or anything" quoted above ).

If Vulkan gets significant traction, then OpenGL will probably follow similar path of being a legacy adapter wrapper around Vulkan at the core. ( Yes Vulkan is "glNext" ( opengl Next) , but from the legacy application perspective that doesn't necessarily need low level control ... it isn't a solution either. ) If Apple can do the same layering with Metal then mac OpenGL can coast pretty far into the future in a very similar state. [ Carbon was around from approx 2000 until 2012 https://en.wikipedia.org/wiki/Carbon_(API) ... that is a pretty long time. I think Apple knows folks need a gradual transition to get off of an entrench legacy API. ]

The more fuzzy future is what happens if Windows+Android+others Vulkan takes off and Apple is standing around doing nothing for an extended period of time. Right now I suspect they are taking the "I'll believe it when I see it" approach. Meanwhile, they can play role that DirectX plays on Windows. Big enough platform ( iOS + tvOS + macOS ) that folks will bother to port to "us" rather than "us" implementing something portable. For games layered on top of gaming engines/library that probably will be true.

Microsoft isn't going to put much effort into maximizing Vulkan on Windows and Google may not execute effectively on Android. Plus Metal is pretty close to Vulkan. Different enough to be more work, but not grossly incompatible either.
 
Well Adobe won't be happy if OpenGL isn't supported since their acceleration features in Photoshop CC require 2.0 or better and Illustrator CC requires 4.0 or better.
 
Couple news from deeps of internet.

RX 480 is locked on stock, 6 pin connector to 1400MHz. At that clock is 5% slower than GTX 1070 with current drivers, that are 4 months old. Final drivers are yet to be delivered to reviewers.

Gaming load power consumption, and compute work(!) at stock clocks does not exceed 110W.

6+8 pin versions of GPUs will not have problem to exceed 1600 MHz mark, and get close to GTX 1080 in performance. We are talking about getting into vicinity of 10-15%. At 1600 MHz GPU stays at 75 degrees celsius, and... 99% GPU load in Battlefield 4, which is DX11 API that underutilized AMD GPUs. So here we have progress. 6+8 pin is not there because of excessive power consumption of OC'ed GPU but rather of circuitry of PCB and the die(no idea, whatsoever...).

Well... HD4850 and HD 4870 times all over again.

General feeling is that AMD took different route this time, and are not competing with Nvidia but creating completely new segment of price and performance.

dec60, now you may know why I said that Nvidia will not be able to compete with AMD in the mainstream segment.
 
Well Adobe won't be happy if OpenGL isn't supported since their acceleration features in Photoshop CC require 2.0 or better and Illustrator CC requires 4.0 or better.

Adobe did a demo at WWDC 2015 State of the Platforms meeting where demo'ed Illustrator and After Effects using Metal and the sky didn't fall in.

It is probably not falling in this year (or next ) either.

Adobe has a portability layer they use to somewhat insulate their apps from platforms. ( If doing both Mac and Windows doesn't want to reinvent the wheel if don't have to. ). OpenGL disappearing from the capabilty layer would be of far bigger concern to Adobe than acceleration. Acceleration hooks tend to be limited and isolated in scope. That is the kind of things that can be ported with reasonable effort. ( Indeed the demo consisted of some stuff that Adobe did in a couple of weeks with early access to Metal. )
 
You know, mago. What if people were digging something that doesn't exist? Just saying.
Maybe Apple will announce it or maybe won't. But at least there's some hope for 6,1.
Asked DarknetGuy, he dont believe the iMac to be ready, the latest he knows its delayed until next year, but still confident on the nMP, about the drivers stated that apple could introduce it on later betas, betas arent closed, Apple use to include and exclude things, until it reach "gloden master" anything could happen...

Ok, so at least polaris 10 support was teased, and Koyot news on RX480 Stock Power consumption at just 110W perfectly match tcMP TDP envelope. I sill hopful on an updated nMP later this summer.

To be earnest, I'll love an earlier update to the iMac 5K and DNG is wrong or a lier, About the tcMP it's clear the best we could see this year is an Xeon e5v4 + Dual RX480 8GB GDDR5X , which is very good for most purposes, and maybe cheapert too, a Mac Pro with dual compute gpu (Vega 10) should cost 2000$ (at least) more than a similar configuration on dual RX480.

So look at next macOS beta's KEXT in case it includes two Polaris 10 Cards, then DNG its rigth and an updated tcMP is on the way.
 
  • Like
Reactions: pat500000
I don't know. I suppose more likely early - mid Q4(October).

still not that bad (RX490 is Vega10?).

I consider Apple should prioritize rMBP, Mac mini (nano) and iMac updates plus the TBD3 release at the same event, and release a partial update tcMP with later silent addition of the RX490 gpu option.

Even Apple should consolidate all the macs as only Macbooks (all retina) trash the Airs, and leave MB12, MB14, MB16 with different options as differentiators, no actual need on an ultra-thin MB14 having a barely thicker MBP14 its irrelevant the differentiation and an unnecessary fragmentation.
 
You are taking my prediction as truth, which may not be really the case. As I have said: I don't know when 490 could come to market.
 
Finally I solved one mystery; why iGPU Iris Pro 5200 has seemed to be "snappier" than Nvidia 750M dGPU on my rMPB; Metal for OS X user interface is working only with Intel iGPU, not Nvidia. This explains a lot of things. I've practically hated Nvidia for this. And OS X that it uses Nvidia, when it thinks I need more power... when iGPU has actually been better. With a second monitor, OS X automatically chooses 750M, and UI starts to feel sluggish.

But on macOS, Apple has finally turned on Metal GUI rendering on Nvidia. So, for me, macOS will bring a long waited feature.
 
Last edited:
Zarni? Isn't it the other way around? Isn't it that Apple updates the drivers for GPUs? You could've installed the Nvidia Web Drivers.

Also, how come was it that Metal was used for HD5200 but not for GT750M? I don't understand.
 
i imagine it will find it's way into your life eventually.. maybe not so much VR at first.
i personally think a well implemented augmented reality would be the sooner technology to be helpful instead of just 'oh wow neat'.

i'd also hazard a guess this is a technology we may see in an apple car.. they could probably do it without the need for the user to don any type of headgear/glasses/etc because with a car, you're already sitting inside a helmet.. a big one.. probably a pretty good environment to be used with VR/AR hardware.
"Apple car" (capitalizing "Apple" is my edit). LOL

How much extra are "rose gold" seat belts?
 
Zarni? Isn't it the other way around? Isn't it that Apple updates the drivers for GPUs? You could've installed the Nvidia Web Drivers.

Also, how come was it that Metal was used for HD5200 but not for GT750M? I don't understand.
Ok, let me put it this way:

Iris Pro is used with laptop screen only. If I open second monitor, OS X automatically starts to use 750M. Computer starts to get hot and GUI is rendered slower.

Also, if there's an app on screen, that has requested hi-performance GPU from OS X, 750M is used, even on single display. And same outcome, sluggish UI, hot computer.

But if I disconnect from power outlet and use battery, iGPU is usually used, and computer feels faster.

But now I know the reason, Apple never turned on the Metal GUI rendering for Nvidia. Not even web-drivers. Even though Metal is supported by Nvidia. GUI rendering with Metal is Intel only and macOS will fix this. Metal on El Captain really is a baby Metal.
 
Last edited:
"Apple car" (capitalizing "Apple" is my edit). LOL

how do you know i wasn't talking about one of these?

26089-Apple-Cars-With-Grape-Wheels.jpg


How much extra are "rose gold" seat belts?
probably a lot


.
.

[doublepost=1466299227][/doublepost]

and Aiden..
the reason i tend to use all lowercase is because i missed school on the day they were teaching capitalization.. i was too busy helping my uncle jack off a horse..

[edit]- i mean my Uncle Jack.. of course.
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.