AMD Announces their GameWorks Equivalent

[zlatan]

Got any documentation on this? Because it contradicts all Kaby Lake info. Or is it just your personal guess?

They said me when I'm ask about their solution to implement a direct support in the engine. Hopefully they will provide some ES chip for testing.

 

[ShintaiDK]

They said me when I'm ask about their solution to implement a direct support in the engine. Hopefully they will provide some ES chip for testing.

So the answer is no, you could just have said so.

 

[zlatan]

So the answer is no, you could just have said so.

Why would I say that when they want to support it?

Honestly, what's the problem with A-sync? This is the de facto standard for variable refresh rate. Ofcourse Intel is looking to support it as fast as they can implement a complete solution around it. This is absolutely normal.

 

[ShintaiDK]

A-Sync is not listed as a feature with Kaby Lake. So unless you can prove other wise there is no merit to your statement.

 

[Dribble]

This is the de facto standard for variable refresh rate.

But it's not - it's supported by a few of the most modern AMD cards, that's it - a tiny % of the total market. Sure it's open and sure Nvidia and Intel one day might support it, but until they do it's not a defacto standard. And yes while it is easy to implement something and call it freesync, it actually takes some effort to create a freesync screen that actually works properly. Hence the fact that most freesync monitors are flawed - there is no quality control. Most of the quality freesync displays are based of gsync displays with their much higher min requirements, and then developed till they work nearly as well as the gsync equivalent.

That's part of the problem with open standards - they tend to be very broad and hence have a very low baseline for quality. Equally no one is in charge, it's no one's job to make sure it reaches some set standard or solves a particular set of needs, or has the right set of development tools, or QA guidelines. This is the case for freesync and will also be the case for any open source amdworks. It's why dev's will continue to use gamesworks - because you get consistent interfaces, code quality, debugging tools and dev support. Not a bunch of random libraries written in different ways with different levels of support and lacking most of the overarching documentation/QA/debug.

 

[zlatan]

But it's not - it's supported by a few of the most modern AMD cards, that's it - a tiny % of the total market. Sure it's open and sure Nvidia and Intel one day might support it, but until they do it's not a defacto standard. And yes while it is easy to implement something and call it freesync, it actually takes some effort to create a freesync screen that actually works properly. Hence the fact that most freesync monitors are flawed - there is no quality control. Most of the quality freesync displays are actually based of gsync displays with their much higher min requirements, and then developed till they work nearly as well as the gsync equivalent.

Why not? I don't see any other standard in the market for variable refresh rate. There is G-Sync, but it's not a standard. Only Asus and Acer support it continuously, which is bad for the technology because they don't able to build as much monitors as it needed to counter the whole A-sync and Freesync fleet. Those two companies will also drop G-sync, because they want to stay profitable, and it is not possible if they have to pay a lot of money to Nvidia. The A-sync standard is a dead end for any proprietary variable refresh rate technology.

That's part of the problem with open standards - they tend to be very broad and hence have a very low baseline for quality. Equally no one is in charge, it's no one's job to make sure it reaches some set standard or solves a particular set of needs, or has the right set of development tools, or QA guidelines. This is the case for freesync and will also be the case for any open source amdworks. It's why dev's will continue to use gamesworks - because you get consistent interfaces, code quality, debugging tools and dev support. Not a bunch of random libraries written in different ways with different levels of support and lacking most of the overarching documentation/QA/debug.

Then who is VESA? They do a lot of QA for any A-sync monitor. AMD do additional QA for the Freesync logo. Intel will do it also.
You probably don't work with GameWorks, or any closed source middleware. The code quality is a mess. Especially for HairWorks. Calculation the geometry for nearly 3-4 ms is a joke. TressFX can do it under 1 ms with higher quality. But the source code is not modifiable, so it is impossible to optimize. Also there is a lot of bugs in the middlewares, and the source code access is allowing me to fix it, and not wait for the middleware provider for weeks.

 

[ShintaiDK]

Adaptive Sync is still an optional part in DP1.3.
http://www.anandtech.com/show/8533/...rt-13-standard-50-more-bandwidth-new-features

Defacto standard, certainly not.

 

[lilltesaito]

A-Sync is not listed as a feature with Kaby Lake. So unless you can prove other wise there is no merit to your statement.

So are you saying that Intel is not going to use A-Sync?

Or are you just trying to muddy the water?

Starting to seem like you just do not want to agree and trying to down play A-Sync because it is something AMD is using. People only have stated that A-Sync is going to be the standard and have showed back up links to this and your only saying that it is just not happen right away, and we all know new tech takes time.

I just want to clear up what you are trying to get at. So if you can answer that first question that would be helpful.

 

[TestKing123]

Yeah it sure does look the same..if you are blind. Sorry but I don't care if NV or AMD looks better or whatever but this is just proofs you are not being objective.

You went through alot of trouble to show up "Nvidia fanboys", but all you did here was show how mindless "AMD fanboys" can be.

 

[ShintaiDK]

So are you saying that Intel is not going to use A-Sync?

I already posted on that matter.

We still have to see if Intel will support it first. It wasn't in Skylake and it wont be in Kaby Lake. That means we are looking at 2018 and counting.

And to note to that, there is no actual commitment done at all at this time.

Or are you just trying to muddy the water?

Starting to seem like you just do not want to agree and trying to down play A-Sync because it is something AMD is using. People only have stated that A-Sync is going to be the standard and have showed back up links to this and your only saying that it is just not happen right away, and we all know new tech takes time.

I just want to clear up what you are trying to get at. So if you can answer that first question that would be helpful.

No, people mudding the water and wanting to create a hyperbole is those that act like its going to be a defacto standard and that its coming in products its certainly not. You can turn your comment around and ask yourself if some people isn't trying to create something that isn't there to promote AMD.

 

[lilltesaito]

I already posted on that matter.



No, people mudding the water and wanting to create a hyperbole is those that act like its going to be a defacto standard and that its coming in products its certainly not.

So you do think that Intel is going to use it, not sure why you can not just say yes.

So now that you agree with everyone else that AMD is not the only one that is going to be using A-Sync, can you answer the following question clearly with a yes or no?

Do you not think that with Intel backing it up(does not matter that they do not have a working product yet since they already publicly said they are going to use it) that it will not be a standard?

 

[ShintaiDK]

So you do think that Intel is going to use it, not sure why you can not just say yes.

So now that you agree with everyone else that AMD is not the only one that is going to be using A-Sync, can you answer the following question clearly with a yes or no?

I never said that. So stop trying to twist the context.

Do you not think that with Intel backing it up(does not matter that they do not have a working product yet since they already publicly said they are going to use it) that it will not be a standard?

Lesson in PR 101. Always follow the money.

Blythe indicated that Intel is positively inclined toward standards-based solutions like Adaptive-Sync, and he said Intel does indeed plan to support this optional extension to the DisplayPort spec. However, Blythe wasn't yet willing to commit to a timetable for Intel enabling Adaptive-Sync support in its products.

This means there is no plans. But just a wait and see game.

 

[Pariah]

Do you not think that with Intel backing it up(does not matter that they do not have a working product yet since they already publicly said they are going to use it) that it will not be a standard?

No, because Intel has no presence in the only market this feature is being used, so what difference does their support make?

 

[lilltesaito]

No, because Intel has no presence in the only market this feature is being used, so what difference does their support make?

From http://techreport.com/news/28865/intel-plans-to-support-vesa-adaptive-sync-displays

Intel's backing would also be a big boost for the Adaptive-Sync standard, since the firm ships by far the largest proportion of PC graphics solutions.

Intel may not be the best for playing games, but they still can be used and having a large pie makes it a no brainier for screen companies to move towards using A-Sync.

 

[Pariah]

Intel may not be the best for playing games, but they still can be used and having a large pie makes it a no brainier for screen companies to move towards using A-Sync.

Ignoring NVidia's 80% market share in the dGPU market in favor of Intel's onboard video is hardly a "no brainier" smart business decision.

 

[Azix]

Ignoring NVidia's 80% market share in the dGPU market in favor of Intel's onboard video is hardly a "no brainier" smart business decision.

nvidia is only at 15% or so of the GPU market.

Considering both intels iGPUs and more so AMDs APUs are perfectly capable of playing the most popular PC games at good settings and high framerates, it wold only be logical to take these things seriously. Intel supporting adaptive sync is a big deal. The fact so many monitor makers are supporting it as well is huge. G-sync has no place in the industry besides making nvidia money at the expensive of gamers. It has no hope of becoming any kind of standard and is just a feature to market. Freesync/adaptive sync is it for us.

 

[flash-gordon]

G-Sync at actual price is not aimed for people who buys low to mid-end GPUs.

nVidia high-end represent how much of this market? That's the share for G-Sync, not 80% of total.

 

[Pariah]

nVidia high-end represent how much of this market? That's the share for G-Sync, not 80% of total.

Who knows, but it's likely a pretty significant percentage of the users that actually know what adaptive Vsync is and care that the products they buy support it.

 

[Paul98]

Very nice, I will be checking out the code and tools when they come out! Hopefully much of it will be useful.

 

[tential]

G-Sync at actual price is not aimed for people who buys low to mid-end GPUs.

nVidia high-end represent how much of this market? That's the share for G-Sync, not 80% of total.

Look how many Freesync monitor choices there are compared to Gsync monitor choices?

It's amazingly weird that many low/mid range users would opt for Nvidia cards, when they can't afford the monitor to go with.

I'm wondering if people will opt for AMD+Budget Freesync options now that they are available quite easily at this point.

AMD's major problem is they haven't really sold the low/mid end crowd on Freesync/Gsync.

 

[sontin]

1) It was WB fault. Signing a deal that doesn't allow to optimize a game is a stupid thing. Nvidia just building a business around the lack of optimization. It makes sense for them.
2) Tessellation is the part of GeometryWorks.
3) AMD provide a same kind of SDK. It's called silhouette tessellation. Nearly five times faster than GeometryWorks with the same quality.

And now you are making things straight up. Wow. :\
"GeometryWorks" was announced one year after Batman:AO and hasnt materialized yet.

Tessellation is part of open standards like DX and OpenGL.

And i hope you can back up your performance numbers. So where can see a fair comparision between AMD's and nVidia's technique?

 

[GodisanAtheist]

Another day another AMD Open initiative.

They need to bring this stuff under one brand/trademark and get it some project development dollars so people can start associating it with AMD/Radeon the way Gameworks is instantly recognizable as an NV initiative.

 

[monstercameron]

Another day another AMD Open initiative.

They need to bring this stuff under one brand/trademark and get it some project development dollars so people can start associating it with AMD/Radeon the way Gameworks is instantly recognizable as an NV initiative.

that is what this thread is about :/

 

[frozentundra123456]

I think he meant AMD itself needs to provide a lot of support and development. Maybe they will. It sounds great for AMD to support "open standards", but somebody needs to take charge, give the project direction and engineering support. Again it reminds me of HSA. I forget the name of the organization, but there is a huge consortium of companies that are "supporting" HSA, but where has it gone?

It is one thing for a company to develop something to completion, and then make it open source for everyone else to use. It is quite another to have some ideas, make them "open source" and hope somebody develops and makes use of it.

 

[GodisanAtheist]

Frozen tundra hit the nail on the head. I feel like we've seen this story before: NV has their closed but highly targeted program that AMD tries to emulate though open standards, a few months go by and AMD's initiative is nowhere to be seen off the rails and in a ditch somewhere because no one was at the helm.

Open standards are great and all but AMD needs to actually provide some real programmers in the trenches to make sure these things are given a direction to grow and thrive.