2016 nMP

[ManuelGomes]

Yep, the axe might come down on someone's head for this. Or not...
I guess Apple could make a custom rTBD that could read this stream instead of DP1.3, to take advantage of the full bandwidth provided by TB3 to enable the 5K HDR mode, and still have the standard DP1.3 mode as an option. But AMD would have to work on their side too on Polaris to enable this, and it's thinking too far off.

 

[MacVidCards]

This thread reminds me that AMD is still healthy enough to keep the PR department's lights on.

I think "Unicorn 3000" is a better name than "Polaris".

 

[Mago]

Yep, the axe might come down on someone's head for this. Or not...
I guess Apple could make a custom rTBD that could read this stream instead of DP1.3, to take advantage of the full bandwidth provided by TB3 to enable the 5K HDR mode, and still have the standard DP1.3 mode as an option. But AMD would have to work on their side too on Polaris to enable this, and it's thinking too far off.

DP 1.3 requires upto 32.5Gbps (inc. overhead), the problem maybe DP1.3 dont allow a single TB channel, only leaves space for usb 3, it actually means a "Thunderbolt Display" actually its only a mere "USB Display", no way to daisychain another TB peripheral, so the pointless of DP 1.3.

But in case HDR and xSync where tooo indispensable, there still the actually more available HDMI 2b.

I dont believe the u-nMP will include DP1.3 (or intel invest on add it to a revised TB3.1), DP 1.2a its far enoug for actual mainstream 4K displays with all the bells on, TB3 its aimed at Mainstream not Specialized hardware aas 5K HDR - Studio, unlikely.

 

[Zarniwoop]

This thread reminds me that AMD is still healthy enough to keep the PR department's lights on.

I think "Unicorn 3000" is a better name than "Polaris".

Here's some more for you, freshly baked out from AMD's pr oven:

http://www.amd.com/en-us/innovations/software-technologies/radeon-polaris#

There're official pictures of both Polaris 11 and 10. More info in those videos embedded.

 

[goMac]

This thread reminds me that AMD is still healthy enough to keep the PR department's lights on.

I think "Unicorn 3000" is a better name than "Polaris".

Well I think it's exciting that another card might be out that doesn't require under clocking for cMPs, and would have a Mac ROM. I don't see why you wouldn't... oh, right.

 

[ManuelGomes]

This was getting too soft, something was missing.. now I know what it was. Damned, quiet was good. Oh well.

Mago, thing is, I don't think Apple can turn back now on HDR really, or maybe they can.

 

[MacVidCards]

This was getting too soft, something was missing.. now I know what it was. Damned, quiet was good. Oh well.

Mago, thing is, I don't think Apple can turn back now on HDR really, or maybe they can.

So an echo chamber full of AMD guys patting each other on the back is a good thing?

Okey doke.

Back to the regularly scheduled commercial programming.

 

[Zarniwoop]

Here's from Intel Fanboys(TM) dept: http://www.anandtech.com/show/10281/intel-adds-crystal-well-skylake-processors-65w-edram

 

[Zarniwoop]

And Nvidia is rumored to introduce Pascal GP104-400 and -200 series (with DICE's Battlefield 5 demos) this week. That should make the Grumpy Old Men dept happy.

This summer should make both Fanclubs happy. Flamethrowers are refueled at the moment. Interesting summer GPU wise.

 

[koyoot]

Yesterday I saw on another forum that someone expects that GTX 1060 Ti will be on GTX 980 Ti level of performance. I laughed as much as I laughed when people expected that Polaris 10 will be on Fury X levels of performance.

Expectations are very high... Its 4 years since we got jump in process nodes.

 

[Mago]

Yesterday I saw on another forum that someone expects that GTX 1060 Ti will be on GTX 980 Ti level of performance. I laughed as much as I laughed when people expected that Polaris 10 will be on Fury X levels of performance.

Expectations are very high... Its 4 years since we got jump in process nodes.

Somebody should explain that mobile users are those which could see a real jump in performance, desktops will only gain efficiency/features from Pascal and Polaris, performance will take off on q2/17 when desktop variants of Pascal 100 and Vega become available.

 

[Zarniwoop]

Somebody should explain that mobile users are those which could see a real jump in performance, desktops will only gain efficiency/features from Pascal and Polaris, performance will take off on q2/17 when desktop variants of Pascal 100 and Vega become available.

True, but with new price points Higher end becomes mainstream, middle class. It will have a big influence on software side (games especially).

Soon everyone have VR ready computer in their hands. Or a computer that has Teraflops of computing power.. also in laptop. It will lead to a new kinds of computers and computational uses.

macOS and next gen apps are probably going to use that extra power in new ways. Looking to you, Siri!

 

[Mago]

True, but with new price point Higher end becomes middle class. It will have a big influence on gaming industry, because current high end graphics become mainstream.

Moreless true, but being realistic don't expect a big performance/$ gain until consumer nVidia GP100 & AMD Vega reach the market in volume.

Just a hint GeForce Titan Z still sells for 3000 and will no face competition until Vega and nVidia Gp100 fills the high end market, from then is when the price cut snowball begins to roll down across the line .

Both nVidia and AMD are market savvy, as they renew the mid range the high end still maintain its value for a while, mid range users either choose the newer mire efficient cards with same performance as the older gen with few new features or buy slightly cheaper (and maybe rebranded ) old gen card with same performance, the high end remains the same, but next year Titan Z will sell for about the same as AMD Radeon pro (fury x2) and fury x2 will cut the half and so on .

 

[koyoot]

I see the debate on other forums about that DX12 will not see any adoption because of numerous reasons.

DX12 was developed by Microsoft with VR in mind. It is exactly why Johanson from DICE was talking with people around the industry about giving the developers more control over hardware if VR is supposed to launch. That is why AMD even first skeptical, started development of it, and after first experiments it was apparent that it will be great thing. From this came Mantle. You may not believe it but it revolved around VR and HSA 2.0, and future specifications of HSA. It had to be open source if it was supposed to be a big hit, and it was: Vulkan and DirectX are gaming VR APIs. LiquidVR and Metal are frameworks for graphical compute, and there will be third API - workstation, CUDA-like, that is revolving around HSA 2.0. And yes, it will also be OpenSource. For the reasons that in another paragraph of this post will be more clear.

I was a bit skeptical about VR. Until I saw demo from Ubisoft. It was simple demo of flying as bird over the city. But it was fantastic idea. And VR in future is everything and everywhere. Cars(!), education, entertainment(films, games, interactive books), HPC, Workstation, home usage, smartphones, computers, even UI. EVERYTHING. This is the market that AMD targeted. Im not saying that this is good thing that one company is going for monopolizing it, but... at least the solutions are Open Source, not proprietary. So if someone will come with better hardware for the software solution - it will always win. And yes, currently for VR, and low-level access to hardware and application better hardware is AMD branded.

Mantle is everywhere currently. Apart from workstation and HPC market. But soon it will be there also. At least a derivative of it .

 

[ManuelGomes]

Are kids allowed in this forum? I guess...

Here's some NVidia PR:
http://wccftech.com/nvidia-geforce-gtx-1080-1070-pascal-teaser/
[doublepost=1462297131][/doublepost]NVidia using the star theme as well?! Come on, where's the imagination...

 

[Mago]

I see the debate on other forums about that DX12 will not see any adoption because of numerous reasons.

DX12 was developed by Microsoft with VR in mind. It is exactly why Johanson from DICE was talking with people around the industry about giving the developers more control over hardware if VR is supposed to launch. That is why AMD even first skeptical, started development of it, and after first experiments it was apparent that it will be great thing. From this came Mantle. You may not believe it but it revolved around VR and HSA 2.0, and future specifications of HSA. It had to be open source if it was supposed to be a big hit, and it was: Vulkan and DirectX are gaming VR APIs. LiquidVR and Metal are frameworks for graphical compute, and there will be third API - workstation, CUDA-like, that is revolving around HSA 2.0. And yes, it will also be OpenSource. For the reasons that in another paragraph of this post will be more clear.

I was a bit skeptical about VR. Until I saw demo from Ubisoft. It was simple demo of flying as bird over the city. But it was fantastic idea. And VR in future is everything and everywhere. Cars(!), education, entertainment(films, games, interactive books), HPC, Workstation, home usage, smartphones, computers, even UI. EVERYTHING. This is the market that AMD targeted. Im not saying that this is good thing that one company is going for monopolizing it, but... at least the solutions are Open Source, not proprietary. So if someone will come with better hardware for the software solution - it will always win. And yes, currently for VR, and low-level access to hardware and application better hardware is AMD branded.

Mantle is everywhere currently. Apart from workstation and HPC market. But soon it will be there also. At least a derivative of it .

MS will do everything in their capabilities to block any initiative similar to HSA, the same way they managed to close (or own) every open std, even if this implies releasing for free their best stuff or poaching staff from open-source world, this has been before and will be the same, MS Windows worst enemy isn't osx, neither linux (as long commercial software don't find a way to be profitable in linux), are the tools that allows migrate apps to rival platforms.

About VR you know what's my opinion,and didn't has changed, still Hype, not a must.

 

[koyoot]

So why Microsoft is putting HLSL(High-Level Shading Language) Open Source? The last bit that was important for graphical Low-Level APIs. Also will be Open Source, and multiplatform. WHOLE industry is currently multiplatform. Why? VR. And simplified development of content, regardless of platform.

HLSL must have been Open Source if VR and HSA will start. It is only to Apple to decide if they want be at least familiar with VR standards or not. Shader Model 6.0 will be very important implementation in Metal.
Here is few helpful links: https://en.wikipedia.org/wiki/High-Level_Shading_Language

 

[Mago]

So why Microsoft is putting HLSL(High-Level Shading Language) Open Source? The last bit that was important for graphical Low-Level APIs. Also will be Open Source, and multiplatform. WHOLE industry is currently multiplatform. Why? VR. And simplified development of content, regardless of platform.

HLSL must have been Open Source if VR and HSA will start. It is only to Apple to decide if they want be at least familiar with VR standards or not. Shader Model 6.0 will be very important implementation in Metal.
Here is few helpful links: https://en.wikipedia.org/wiki/High-Level_Shading_Language

For the same reason , to block any open-source related initiative, this gives they a chance to ready an improved proprietary alternative but compatible with but most important: Windows/XboX only.

 

[ManuelGomes]

Looks like BDW-E is a tad (10%+) faster than HSW-E, this should apply to -EP as well.
OC seems somewhat better too, if that's your thing. Not in a Mac though

Mago, I'm with you man, this VR sudden fever gets on my nerves. Well, there has to be something that people really think they need to justify buying new hardware I guess.

 

[koyoot]

For the same reason , to block any open-source related initiative, this gives they a chance to ready an improved proprietary alternative but compatible with but most important: Windows/XboX only.

Look at their comments on HLSL Open Source. It has nothing to do with situation you are describing.

 

[Mago]

Look at their comments on HLSL Open Source. It has nothing to do with situation you are describing.

It's just a Jump forward tactic, you don't show me the angel, I lived in the inferno I know Satan face.

You'll see some time from now something new from ms that makes open HLSL irrelevant .

 

[ManuelGomes]

4th beta OS X 10.11.5 for developers, they're coming out strong...

 

[koyoot]

It's just a Jump forward tactic, you don't show me the angel, I lived in the inferno I know Satan face.

You'll see some time from now something new from ms that makes open HLSL irrelevant .

They would make whole Shader Model specification irrelevant? HLSL is very important part of it. Whole Microsoft film was revolving about Direct3D 12, and HLSL and Shader Model 6.0.

I loved the first paragraph of your post .

I was writing last post about VR, and its impact on industry. I looked then at iPad Pro and iPhone 6S graphical performance benchmarks. It made even more sense...

 

[Mago]

They would make whole Shader Model specification irrelevant? HLSL is very important part of it. Whole Microsoft film was revolving about Direct3D 12, and HLSL and Shader Model 6.0.

I loved the first paragraph of your post .

I was writing last post about VR, and its impact on industry. I looked then at iPad Pro and iPhone 6S graphical performance benchmarks. It made even more sense...

Not the whole specification but the products based on "pure std HLSL" not they 'MS Souped Up HLSL" and this will happen much earlier than the industry can revamp HLSL to reach the new functionality or performance of MS suHLSL


Ah, I have a new fav phrase "next time I have an Idea like this, punch me in my face - Tyrion Lannister"

 

[koyoot]

It will not happen. Simply because HLSL is Open Source. Nobody can block it. If Microsoft will come with something better, proprietary, nobody may want to use it. Why? Because it will be proprietary. We are not in mid 90's where everybody is relying on MS. And that would be complete opposite why Low-Level APIs came to live - to give developers biggest control over hardware. Thats what they have asked for. Open Source, multiplatform option that is platform agnostic. It was the only possible way VR in current idea and form could be done. Microsoft sees this, and they are not stupid.