Crimson 16.9.1 reduces CPU overhead by quite a margin, several titles see boots in performance.

[IllogicalGlory]

Would you mind posting some links. My google fu must suck.

http://youtu.be/VDo-j00vUtw

Digital Foundry testing Sandy to Skylake i7s at stock. 6700K is 20% faster in GTAV and 20-25% faster in Far Cry 4. It also does well in Witcher and COD:AW. Skylake looks like very good in these tests. I do somewhat regret getting a 4790K, though the price was right at the time.

 

[BFG10K]

This is good to see. Despite some living in a fantasy world of denial, AMD's drivers have significantly more CPU overhead than nVidia's. This becomes even more apparent for those on a budget who buy slower CPUs and they need to eke as much performance as possible.

AMD can't rely on Metal APIs as a crutch. Mantle is dead, and DX12/Vulkan games will always be a drop in the bucket compared to DX11 titles. Hell, there are hundreds of new indie titles that are still targeting DX9/XP.

So again, AMD removing driver overhead is good for customers and good for competition.

 

[Alqoxzt]

Can someone please tell me what does surface format optimization option do in terms of graphics and performance?

 

[richaron]

AMD can't rely on Metal APIs as a crutch. Mantle is dead, and DX12/Vulkan games will always be a drop in the bucket compared to DX11 titles. Hell, there are hundreds of new indie titles that are still targeting DX9/XP.

This is the kind of thinking which got us into this mess in the first place.

I like the way you slipped in a backhanded compliment, almost as if posting an objective comment, and then immediately resorted with this myopic anti-AMD fud.

I'll agree lower DX11 overhead is a nice achievement. But I'll have to correct you; Mantle has inspired and evolved into two of the most radical graphics APIs in history. And more importantly, to claim DX12 will be "always be a drop in the bucket" compared to DX11 is frankly stupid. It's a willingly shortsighted claim (unless you really mean forever?), and since people come to forums like this for information they need to know this.

 

[BFG10K]

I like the way you slipped in a backhanded compliment, almost as if posting an objective comment, and then immediately resorted with this myopic anti-AMD fud.

You sound angry. Why is that?

But I'll have to correct you; Mantle has inspired and evolved into two of the most radical graphics APIs in history.

Please stop spreading misinformation. DX12 was in development years before Mantle even existed.

As for "radical", the industry used to code to the metal years ago. It was called DOS and Assembler. Was DOS "radical"? How about x86 assembler?

And more importantly, to claim DX12 will be "always be a drop in the bucket" compared to DX11 is frankly stupid. It's a willingly shortsighted claim (unless you really mean forever?), and since people come to forums like this for information they need to know this.

On Windows DX12 will gain about as much traction as PhysX did, Vulkan even less so. Generally speaking only the biggest AAA engine houses will be able to afford the resources to properly code for DX12. Windows has thousands of games released for it every year, including huge numbers of indie games written using garbage-collected C#/.NET. The industry has been moving away from low level APIs for years.

Even if we take what we have so far, about half of DX12 games have been a shocking failure, running slower and/or buggier than DX11. The mantra "things will get better" doesn't hold given the hardware is ever changing, and DX12 paths will need constant app-side patching to maintain performance advantages for new hardware. Almost every single Mantle game now runs slower than DX11 on newer AMD hardware, yet it's quietly swept under the rug.

 

[linkgoron]

Hell, there are hundreds of new indie titles that are still targeting DX9/XP.

While I generally agree with your sentiment (that AMD needs to fix their overhead, if it's possible), I completely disagree with this statement. These "hundreds" of new indie titles probably easily run at 60+ FPS on an Rx 470/480 (as I wouldn't expect a low-end user to have a 144hz Freesync monitor). It's the AAA titles that need the CPU overhead fixed, not barely-demanding DX9 indie games.

 

[BFG10K]

These "hundreds" of new indie titles probably easily run at 60+ FPS on an Rx 470/480 (as I wouldn't expect a low-end user to have a 144hz Freesync monitor). It's the AAA titles that need the CPU overhead fixed, not barely-demanding DX9 indie games.

I agree, and that's my point. Anyone expecting DX12 to proliferate Windows gaming is living in a fantasy world, and your's is another example why that won't happen.

Like I said above, generally speaking it'll only be the biggest AAA vendors, so it'll remain a niche API in the grand scheme of things.

 

[Carfax83]

Like I said above, generally speaking it'll only be the biggest AAA vendors, so it'll remain a niche API in the grand scheme of things.

That's how it should be though. I'm pretty sure Microsoft never intended DX12 to be a mainstream API, as low level APIs are much harder to program for; especially on an open platform like the PC.

For indie and more mainstream developers, DX11.3 should suffice. Only elite developers should mess with DX12 and Vulkan.

 

[Carfax83]

Please stop spreading misinformation. DX12 was in development years before Mantle even existed.

When is this myth ever going to die? It makes no sense for DX12 to be derived from Mantle, as the timeline doesn't allow for it.

 

[Red Hawk]

Please stop spreading misinformation. DX12 was in development years before Mantle even existed.

As for "radical", the industry used to code to the metal years ago. It was called DOS and Assembler. Was DOS "radical"? How about x86 assembler?

You please stop spreading misinformation. Mantle came out before DX12; how exactly would DX12 have been in development before Mantle "even existed"?

I think people sometimes exaggerate how much Mantle contributed to to DX12, code-wise. We have clear evidence that the work on Mantle transferred directly over to Vulkan; not so much with DX12. Still, I think it's pretty clear that Mantle had influence on DX12 by being the first modern low-level API. If Mantle hadn't gone in that direction, I doubt DX12 would have. And yes, it was "radical" (I think "revolutionary" is a better word though). So what if DOS and Assembler were used to code close to the metal years ago (more like decades when it comes to games)? We're talking about modern graphics APIs, and before Mantle came along the norm with graphics APIs from Microsoft and Khronos was to be high-level with thick driver layers. It's because of Mantle that both companies moved towards low-level (and no coincidence that Apple created their own low-level API, Metal). It's' a sea change, one that was started by Mantle.

 

[Glo.]

AMD themselves claimed that Mantle spawned in one way or another all of the low-level APIs that are currently in industry. Mantle is directly in Vulkan and LiquidVR. In DX12 and Metal - indirectly. Feature set for all of those APIs comes from Mantle, however.

https://community.amd.com/community/gaming/blog/2015/05/12/on-apis-and-the-future-of-mantle

1. Mantle must take on new capabilities and evolve beyond mastery of the draw call. It will continue to serve AMD as a graphics innovation platform available to select partners with custom needs.
2. The Mantle SDK also remains available to partners who register in this co-development and evaluation program. However, if you are a developer interested in Mantle "1.0" functionality, we suggest that you focus your attention on DirectX® 12 or GLnext.

What this means directly what AMD sais is that Mantle 1.0 is in DirectX 12. So misinformation actually is saying that Mantle is not in DX12. It is. And GLNext is Vulkan.

 

[Janooo]

...
Please stop spreading misinformation. DX12 was in development years before Mantle even existed.
...

What an irony, this is such a fat misinformation!!!
You are spreading them.

 

[nurturedhate]

DX11 will never be more than niche, long live DX10. DX10 will never be more than niche, long live DX9. We've heard all of this before, over and over again. Want to keep going?

 

[daperl]

Like I said above, generally speaking it'll only be the biggest AAA vendors, so it'll remain a niche API in the grand scheme of things.

This statement seems to be a contradiction.

 

[Carfax83]

Still, I think it's pretty clear that Mantle had influence on DX12 by being the first modern low-level API. If Mantle hadn't gone in that direction, I doubt DX12 would have.

For this to be true, you have to believe that low level APIs are created in mere months, and not years. Lets look at the timeline shall we:

1) AMD announces Mantle on September 26 2013.
2) Mantle enabling driver for BF4 becomes available on Feb 3 2014
3) Microsoft announces DX12 on March 20 2014, and shows Forza 5 running on a Titan Black using DX12
4) NVidia blog posted on March 20 2014 claims that work on DX12 started more than 4 years prior, and that for the past year NVidia had been working with Microsoft to deliver a working implementation for presentation at GDC

Now lets put things into context. AMD Mantle specifically targeted Radeon hardware. DX12 on the other hand supports a much wider range of hardware, to not only include both Radeon and Geforce, but phones, tablets, laptops etcetera as well.. If Mantle influenced DX12, then how and why was the first DX12 demo run on NVidia hardware, a mere month after Mantle first became available to consumers?

DX12 is much more complex than Mantle by virtue of the various hardwares that it supports and so had a much longer development time and greater collaboration between IHVs and Microsoft.. So, how on Earth can you believe Mantle had anything to do with DX12? They are two different beasts entirely. The only thing they have in common are that they are low level APIs.

In fact, AMD posted in their own Mantle FAQ page that there was no relationship between Mantle and DX12..

So let's put this myth to bed now shall we?

 

[Glo.]

For this to be true, you have to believe that low level APIs are created in mere months, and not years. Lets look at the timeline shall we:

1) AMD announces Mantle on September 26 2013.
2) Mantle enabling driver for BF4 becomes available on Feb 3 2014
3) Microsoft announces DX12 on March 20 2014, and shows Forza 5 running on a Titan Black using DX12
4) NVidia blog posted on March 20 2014 claims that work on DX12 started more than 4 years prior, and that for the past year NVidia had been working with Microsoft to deliver a working implementation for presentation at GDC

Now lets put things into context. AMD Mantle specifically targeted Radeon hardware. DX12 on the other hand supports a much wider range of hardware, to not only include both Radeon and Geforce, but phones, tablets, laptops etcetera as well.. If Mantle influenced DX12, then how and why was the first DX12 demo run on NVidia hardware, a mere month after Mantle first became available to consumers?

DX12 is much more complex than Mantle by virtue of the various hardwares that it supports and so had a much longer development time and greater collaboration between IHVs and Microsoft.. So, how on Earth can you believe Mantle had anything to do with DX12? They are two different beasts entirely. The only thing they have in common are that they are low level APIs.

In fact, AMD posted in their own Mantle FAQ page that there was no relationship between Mantle and DX12..

So let's put this myth to bed now shall we?

You should read the AMD blog post about Mantle 1.0 functionality in DX12, and OpenGLNext(Vulkan).

It is in post number 36 in this thread, on this very page. I will link it for ya, after all: https://forums.anandtech.com/thread...s-in-performance.2486639/page-2#post-38476182

Have you actually thought that MS could have been developing DX12 for quite some time, but without addition of Mantle it would not be low-level API?

You know why is that? Because Microsoft has something like DX11.3 which has most of features of DX12. Apart from the low-level ones. And all of the features that are missing from DX11.3, are somewhat familiar to those who worked with Mantle feature set.

So no, DirectX12 really has Mantle feature set/code inside it.

 

[Red Hawk]

For this to be true, you have to believe that low level APIs are created in mere months, and not years.

And you seem to believe that MS and AMD didn't know about Mantle and DX12 respectively before the general public did. I highly doubt that. APIs are collaborative efforts, with input from the OS maker, IHVs, and software/game developers. Mantle could have influenced the direction of DX12 while both were still in development, causing DX12 to change course to be more low-level. I find that more plausible than AMD putting resources into an exclusive API when Microsoft was going to come out with a low level API anyways just a year or so later.

In fact, AMD posted in their own Mantle FAQ page that there was no relationship between Mantle and DX12..

So let's put this myth to bed now shall we?

No direct relationship. Stop distorting your sources to say what you want them to say.

 

[Carfax83]

You should read the AMD blog post about Mantle 1.0 functionality in DX12, and OpenGLNext(Vulkan).

It is in post number 36 in this thread, on this very page. I will link it for ya, after all: https://forums.anandtech.com/thread...s-in-performance.2486639/page-2#post-38476182

Read the post, and I see no evidence of any connection with DX12.

Have you actually thought that MS could have been developing DX12 for quite some time, but without addition of Mantle it would not be low-level API?

Speculation is fun, but without any evidence whatsoever it becomes quite dull..

You know why is that? Because Microsoft has something like DX11.3 which has most of features of DX12. Apart from the low-level ones. And all of the features that are missing from DX11.3, are somewhat familiar to those who worked with Mantle feature set.

So no, DirectX12 really has Mantle feature set/code inside it.

The main purpose of low level APIs is to remove CPU overhead and increase rendering parallelism. That's where most of the performance increase comes from with both DX12 and Vulkan. You guys need to get over the idea that only Mantle can be a low level API, much like some of you think that only AMD can do asynchronous compute.

Microsoft is one of the largest tech companies in the World, and actually have the resources, capability and experience to develop whatever kind of API they want without the need for AMD's input; although to be sure, AMD definitely contributed greatly to DX12, along with NVidia, Intel, Qualcomm and many other IHVs..

If there is any connection between DX12 and Mantle, then AMD might have contributed some Mantle code during DX12's development on their behalf. But this doesn't mean that DX12 was inspired by, or created from Mantle.

 

[Carfax83]

And you seem to believe that MS and AMD didn't know about Mantle and DX12 respectively before the general public did. I highly doubt that. APIs are collaborative efforts, with input from the OS maker, IHVs, and software/game developers. Mantle could have influenced the direction of DX12 while both were still in development, causing DX12 to change course to be more low-level. I find that more plausible than AMD putting resources into an exclusive API when Microsoft was going to come out with a low level API anyways just a year or so later.

This is plausible, but as I said to Glo. above, it lacks any evidence and without evidence, it's just speculation. I posted a link to an AMD Mantle FAQ page which stated that there is no direct relationship between Mantle and DX12. That's straight from the horse's mouth. What more do you need?

No direct relationship. Stop distorting your sources to say what you want them to say.

What does it matter? The implication is still the same regardless. DX12 and Mantle have no connection at all. And why would they, when DX12 supports multiple IHVs and architectures, whilst Mantle supports only one IHV, and one specific architecture; GCN.

 

[flopper]

Mantle and dx12 shared the same doc, same sentences, same info, if dx12 had been in development so long I am sure they would have a whole documentation that been vastly different.
It wasnt.
cant trust Nvidia when they say things, you know that right?

 

[Bryf50]

It's like the people in this thread live in a world where common game engines don't exist and 2D indie platformers are the main reason people upgrade their GPUs.

 

[Zstream]

I'm SHOCKED that BFG retorted to the old statement that AMD did not impact DX12 development. Why do you think Mantel or DX12 started in the first place? They could have easily have used 11.3 specs and be done. The issue comes down to consoles and AMD pushing console low level work to PC's.

 

[Red Hawk]

This is plausible, but as I said to Glo. above, it lacks any evidence and without evidence, it's just speculation. I posted a link to an AMD Mantle FAQ page which stated that there is no direct relationship between Mantle and DX12. That's straight from the horse's mouth. What more do you need?



What does it matter? The implication is still the same regardless. DX12 and Mantle have no connection at all. And why would they, when DX12 supports multiple IHVs and architectures, whilst Mantle supports only one IHV, and one specific architecture; GCN.

I'll admit there's no explicit evidence of how Mantle influenced DX12's development. I just think the context of AMD working with Microsoft on the Xbox One, pushing for a low-level API first on PC and Microsoft following suit with a similar feature set in DX12 points to AMD and Mantle having a significant influence on DX12. I'd prefer to describe it as "influenced" rather than "inspired", but in either case, it's not some kind of myth. More like an educated guess. And AMD gets credit for going first and being the trendsetter.

What is a myth is to say that DX12 was directly derived from/is Mantle. That's pretty baseless, but I don't think anyone in this thread was saying that.

And no, "no direct relationship" is not the same thing as "no connection at all". Again, please stop distorting AMD's statement.

 

[master_shake_]

You sound angry. Why is that?


Please stop spreading misinformation. DX12 was in development years before Mantle even existed.

As for "radical", the industry used to code to the metal years ago. It was called DOS and Assembler. Was DOS "radical"? How about x86 assembler?


On Windows DX12 will gain about as much traction as PhysX did, Vulkan even less so. Generally speaking only the biggest AAA engine houses will be able to afford the resources to properly code for DX12. Windows has thousands of games released for it every year, including huge numbers of indie games written using garbage-collected C#/.NET. The industry has been moving away from low level APIs for years.

Even if we take what we have so far, about half of DX12 games have been a shocking failure, running slower and/or buggier than DX11. The mantra "things will get better" doesn't hold given the hardware is ever changing, and DX12 paths will need constant app-side patching to maintain performance advantages for new hardware. Almost every single Mantle game now runs slower than DX11 on newer AMD hardware, yet it's quietly swept under the rug.

We’ve spoken to several sources with additional information on the topic who have told us that Microsoft’s interest in developing a new API is a recent phenomenon, and that the new DirectX (likely DirectX 12) will substantially duplicate the capabilities of AMD’s Mantle. The two APIs won’t be identical — Microsoft is doing its own implementation — but the end result, for consumers, should be the same: lower CPU overhead and better scaling in modern titles.

http://www.extremetech.com/gaming/1...rectx-12-will-imitate-and-destroy-amds-mantle
http://www.extremetech.com/gaming/1...rectx-12-will-imitate-and-destroy-amds-mantle
just saying.

 

[Carfax83]

Mantle and dx12 shared the same doc, same sentences, same info, if dx12 had been in development so long I am sure they would have a whole documentation that been vastly different.
It wasnt.

Where's your evidence?