Are you dissapointed by DX12?

Stoly

Supreme [H]ardness
Joined
Jul 26, 2005
Messages
6,713
As far as I remember, every new DX release has been plagued with issues. So I'm not that surprised that DX12 is no different.

But with the exception of GoW, pretty much every other DX12 game has issues, even more so on nvidia hardware.

I think its about time Devs and GPU builders do something to revert the situation, otherwise it looks like DX10 all over again.
 
Nah. Only thing that's a downside for me is that devs have to include multi-GPU support, so my second card is likely to be used less and less with newer titles.

The rest is just teething issues, it'll work itself out as it always does.
 
No because this is not my first rodeo and I didn't swallow the hype. When it works great and is reliable then I will care. Until then, it is vaporware like so much else.
 
It'll still get better over time, but yeah people hype it too much. Think about the logic: Game developers, who repeatedly push out games with massive issues and rely on being able to fix them later on via updates, have so much free time that with lower level access to hardware they're going to optimize their games on every piece of hardware and do a better job than the the hardware manufacturers can. Yeah eventually, but unless you have tons of resources, which who does, it'll be awhile.
 
As far as I remember, every new DX release has been plagued with issues. So I'm not that surprised that DX12 is no different.

But with the exception of GoW, pretty much every other DX12 game has issues, even more so on nvidia hardware.

I think its about time Devs and GPU builders do something to revert the situation, otherwise it looks like DX10 all over again.

The issue with DX 10 was the adoption rate of Vista, 10 is already well past that for gamers if Steam's Hardware Survey is even remotely accurate. It's still a new API, it'll take some time for it to mature as is always the case. It took time for DX 11 to replace DX 9. As for games that have done DX 12 well, there's also Forza Horizon 3. One can debate if it's as optimized as GoW 4 but FH3 has been extremely well received and it's a great racer.
 
After Vulkan I'm far less impressed. Seems more of an OS push than a performance reason.

Vulkan support seems like the way to go. When DX12 works, I usually benefit from it. But Doom set a new bar and DX12 hasn't reached it yet. I think both are good for gaming as a whole.
 
It has yet to provide the kind of speed improvements initial reports indicated could be possible. I think it will take more time developers to figure out the best way to use it.
 
DX12 is a console API, not a PC API. And we have seen just the first few problems of it, but the future will hold much more as games isn't updated. Its worse than DX10 ever was.

For Intel DX12 have been a disaster, not because they dont support DX12. But because the developers doesn't include it. Civ6 is a game perfectly capable on the IGP, yet only DX11 support there.
 
Forza 3 Horizon looks really good and runs just as well on my PC. It's DX12 also and has settings that far exceed what my current setup is capable of and maintain 60fps. If compared strictly on system performance to Project Cars, Forza 3 Horizon is a godsend. Completely blows the doors off Project Cars at all visual settings. So I can't say that DX12 has been a disappointment. Of course we have the titles that give zero benefits for Nvidia under DX12 and even a few that do the same for AMD.

Even though it's in it's infancy, we should set the bar for Doom performance when communicating with game developers on their forums. Just tell them that there is a standard set by ID Software and we expect greatness from you also. No need to berate them; just give them your honest expectations. We have money in our pockets and you want it. This is what we expect. Email me some marketing materials when you achieve this level of smoothness.
 
DX12 is a console API, not a PC API. And we have seen just the first few problems of it, but the future will hold much more as games isn't updated. Its worse than DX10 ever was.

For Intel DX12 have been a disaster, not because they dont support DX12. But because the developers doesn't include it. Civ6 is a game perfectly capable on the IGP, yet only DX11 support there.

Is that why there is an option to run Civ6 in DX12 mode?

This thread is just a lot of nothing. It usually takes awhile for developers to include the latest APIs into their games. There have been a number of improvements from DirectX in the last few releases, but all of them have taken awhile for developers to adapt. I would say the same will happen with DX12. After another year or two, then maybe we could see some judgement on its actual usefulness.
 
yes very much disappointed in DX12...seems like DX11 runs better and even has better exclusive features such as VXAO
 
yes very much disappointed in DX12...seems like DX11 runs better and even has better exclusive features such as VXAO

I think perhaps you are confused here? VXAO is not an included feature of DX11, its an Nvidia feature. It is also not technically limited to DX11, it just hasn't been implemented anywhere with DX12 yet. There are many opinions on that matter atm whether it will be supported by games in the future with DX12. As far as I know, all the features included with DX11 are also included with DX12.

I think my apprehension of DX12 is well known by now.

Vulkan is the way to go, but then again, once upon a time so was OpenGL.

I would like to see the adoption of more APIs myself. DX hasn't had much competition forever it seems, OpenGL never seems to get out of its own way. It remains to be seen how much Vulkan will push DX, personally I hope there is more adoption of it over DX12, just so we can get some more competition there. Although I am not convinced that Microsoft will continue to support DX as much if Vulkan proves to be better, they may just adopt/push Vulkan instead.
 
Is that why there is an option to run Civ6 in DX12 mode?

This thread is just a lot of nothing. It usually takes awhile for developers to include the latest APIs into their games. There have been a number of improvements from DirectX in the last few releases, but all of them have taken awhile for developers to adapt. I would say the same will happen with DX12. After another year or two, then maybe we could see some judgement on its actual usefulness.

You can try run Civ6 in DX12 on a fully capable DX12 IGP and it simply crashes. Funny considering the IGP is the most feature complete of all 3 IHVs. Its also another statement that DX12 support have to be done per uarch type, even down to SKU level. Vulkan is no different either.
 
You can try run Civ6 in DX12 on a fully capable DX12 IGP and it simply crashes. Funny considering the IGP is the most feature complete of all 3 IHVs. Its also another statement that DX12 support have to be done per uarch type, even down to SKU level. Vulkan is no different either.

I have run Civ6 with Dx12, it did happen to crash, but not until after about 2 hours of multiplayer gameplay. The point is, Civ6 is offering support for DX12, and it really isn't DX12's fault that it might not be the best implementation. The developers need to learn how to program the games to best use DX12. Blaming an API because developers haven't fully learned to use it yet is a bit premature. Typically games need to be designed from the ground up to use certain APIs, not hack up some code at the last minute to try and use it just because its there. Even Microsoft wasn't expecting any games to have adopted DX12 until December of this year. And that is just preliminary games. Given development lifecycles of games these days, we are not likely to see many AAA games properly implementing DX12 until next year on.
 
...
Its also another statement that DX12 support have to be done per uarch type, even down to SKU level. Vulkan is no different either.

While this statement is true, most developers I know aren't bothered about whether their code favors an IHV or not. Most developers are more vendor neutral than people would like to believe. I know I hate both Nvidia and AMD equally due to the grief they give me.

That said, I do get a lot of issues testing DX12 on Intel's IGPs. I don't know why, suspect Intel is not spending enough resources on their drivers or something.
 
While this statement is true, most developers I know aren't bothered about whether their code favors an IHV or not. Most developers are more vendor neutral than people would like to believe. I know I hate both Nvidia and AMD equally due to the grief they give me.

That said, I do get a lot of issues testing DX12 on Intel's IGPs. I don't know why, suspect Intel is not spending enough resources on their drivers or something.
Which is exactly the issue, isn't it? That you need to be vendor-specific to get the most out of DX12.
 
Which is exactly the issue, isn't it? That you need to be vendor-specific to get the most out of DX12.

Yes and Vulkan doesn't solve the problem either. OpenGL also had this problem in a way via extensions.

At least Vulkan and OpenGL is supported on more platforms. That itself is a huge selling point to us. With Vulkan being an API of choice for Android now, there is hope that Vulkan is the foundation of the NVN API for the Nintendo Switch.

I can dream, can I?
 
I have run Civ6 with Dx12, it did happen to crash, but not until after about 2 hours of multiplayer gameplay. The point is, Civ6 is offering support for DX12, and it really isn't DX12's fault that it might not be the best implementation. The developers need to learn how to program the games to best use DX12. Blaming an API because developers haven't fully learned to use it yet is a bit premature. Typically games need to be designed from the ground up to use certain APIs, not hack up some code at the last minute to try and use it just because its there. Even Microsoft wasn't expecting any games to have adopted DX12 until December of this year. And that is just preliminary games. Given development lifecycles of games these days, we are not likely to see many AAA games properly implementing DX12 until next year on.

Did you ignore the IGP part? It doesn't run at all in DX12 on it because the developers didn't include a specific DX12 path for it.

Now take a wild guess what happens when you try and play old games on (radically) new hardware and you dont have something like a DX11 option to fall back on. Because I doubt the developers will update old games for their blue eyes sake. You can call it planned obsolescence of games unless you like to collect old hardware.
 
DX12 is meh. I felt taking away performance in multi-gpus isn't moving PC gaming forward.
Make it more efficient would have been ideal. Visually it's the same as DX11.
 
Nah. Only thing that's a downside for me is that devs have to include multi-GPU support, so my second card is likely to be used less and less with newer titles.

The rest is just teething issues, it'll work itself out as it always does.
this....i really want dual gpus to actually to be useful!!!!! That was the biggest thing about DX12. high res textures and multi gpu. Still almost no reason to use GPUs unless you play AAAA games like COD/BF. TT
 
Did you ignore the IGP part? It doesn't run at all in DX12 on it because the developers didn't include a specific DX12 path for it.

Now take a wild guess what happens when you try and play old games on (radically) new hardware and you dont have something like a DX11 option to fall back on. Because I doubt the developers will update old games for their blue eyes sake. You can call it planned obsolescence of games unless you like to collect old hardware.

I guess I just don't see your issue then..? So Civ6 DX12 will not work on IGP. But you can play Civ6 DX11 on IGP. If the problem persists in the future and either the developer or Intel does not fix the capabilities of their IGP with DX12, then I guess that is an issue for them. I fail to see how that is DX12's issue? You keep bringing up things that are not inherently problems for the API, but for those developing their products to use the API.
 
Which is exactly the issue, isn't it? That you need to be vendor-specific to get the most out of DX12.
vendor specific? It just shows Intel has shitty drivers..... The day intel actually beefs iGPUs and has good drivers will be awesome....one day.....
 
vendor specific? It just shows Intel has shitty drivers..... The day intel actually beefs iGPUs and has good drivers will be awesome....one day.....

Yes, as part of DX12, the reason it is supposed to help boost speed, efficiency, and allow multi-gpu support is they opened up the lower level calls to hardware. This is in lieu of how they did it before where the hardware layer was abstracted and all those lower calls were handled by the OS/API. Therefore, in order to take advantage of the new features in DX12, developers actually have to write code specific to the hardware, which typically means specific to the particular vendor.
 
this....i really want dual gpus to actually to be useful!!!!! That was the biggest thing about DX12. high res textures and multi gpu. Still almost no reason to use GPUs unless you play AAAA games like COD/BF. TT

I truly believe mGPU died at the starting line. Apart from it being a technical exercise for developers who want a challenge, I don't see how any beancounter will greenlight mGPU support.
 
I guess I just don't see your issue then..? So Civ6 DX12 will not work on IGP. But you can play Civ6 DX11 on IGP. If the problem persists in the future and either the developer or Intel does not fix the capabilities of their IGP with DX12, then I guess that is an issue for them. I fail to see how that is DX12's issue? You keep bringing up things that are not inherently problems for the API, but for those developing their products to use the API.

We know DX12 works on the IGP.

The problem is the developer doesn't make a path for it. What do you think happens when you got a DX12 only game and you replace your Titans with say Volta or further out. And suddenly it may not work, depending on the will of the developers to patch old games? Going to keep one of those older Titans as a solution?

We got a good example with Mantle and BF how it quickly goes. And that was with tiny GCN changes staying in the same uarch.
 
We know DX12 works on the IGP.

The problem is the developer doesn't make a path for it. What do you think happens when you got a DX12 only game and you replace your Titans with say Volta. And suddenly it may not work? Going to keep one of those older Titans as a solution?

What happens when you keep making up obscure potential problems based off of one example? How is that particular problem being solved by any of the other competing APIs? They all have some similar issues. And while DX12 allows you to make lower level calls, it isn't necessarily forcing you to. So what is your point here? Companies are going to code to be completely specific to one vendor and all hell is going to break loose? That is on that company, not the API.
 
DX12 is the nonstarter that anyone paying attention had seen coming from miles away. And now that Windows 10 uptake has stalled, even less reason for developers to bother with a 10-only API, since Windows 7 and 8.x combined make up more than half the market, and will remain there for years until Microsoft returns to sanity with a proper Windows successor that respects users and drops the fischer-price phone UI and app bullshit.

id set the more sensible bar with Vulkan for DOOM, now hopefully more developers will wake up to the realization that with Vulkan they can have all the multi-platform, multi-Windows-versions upsides of DX12 with none of the technical or financial downsides.
 
Last edited:
I truly believe mGPU died at the starting line. Apart from it being a technical exercise for developers who want a challenge, I don't see how any beancounter will greenlight mGPU support.

Exactly, people have to think on the money before their dreams. Unlike Nvidia and AMD, developers have no interest in mGPU since it doesn't increase their sale at all.
 
What happens when you keep making up obscure potential problems based off of one example? How is that particular problem being solved by any of the other competing APIs? They all have some similar issues. And while DX12 allows you to make lower level calls, it isn't necessarily forcing you to. So what is your point here? Companies are going to code to be completely specific to one vendor and all hell is going to break loose? That is on that company, not the API.

Its not an issue with a high level API.

See, you put the burden on the developer. But you forget the developer isn't going to care. After a year or maybe longer if lucky, the developers have moved on. You are asking the developer to spend extra money, resources and time on a game that already sold and wont sell more as such. Just to spend a lot of work to get it to work on future graphic architectures. Unlike the IHV who have an interest in making it work.

And yes, this is exactly what DX12 is or Vulkan for that matter. Its flawed by design. Great for consoles, disaster on PC.
 
What happens when you keep making up obscure potential problems based off of one example? How is that particular problem being solved by any of the other competing APIs? They all have some similar issues. And while DX12 allows you to make lower level calls, it isn't necessarily forcing you to. So what is your point here? Companies are going to code to be completely specific to one vendor and all hell is going to break loose? That is on that company, not the API.

I was trying to explain this to our summer intern tester.

Pretend you want a cheeseburger meal.
I can tell my intern: "I want a cheeseburger meal, go get me one!". That's my API call. No matter whether he goes to In-n-Out or McDonalds, I get a cheeseburger meal. We can start to nitpick about the cheeseburger from In-n-Out or McDonalds, like its cheese, or ketchup, or onions, or burger patty(image quality), but it's still a cheeseburger.

That's DX11.

Now with DX12, I have to go buy my own. I can go to either McDonalds or In-n-Out on my own to get the cheeseburger. Now, if I want to, McDonalds or In-n-Out can even let me customize my own cheeseburger! No pickles? No problem! Salad cream instead of ketchup? It's your poison.....

Then, you can walk into a KFC, which doesn't have cheeseburgers.....


Edit: I know it's a terrible analogy, but it was funny when I tried it on my intern. :p
 
Its not an issue with a high level API.

See, you put the burden on the developer. But you forget the developer isn't going to care. After a year or maybe longer if lucky, the developers have moved on. You are asking the developer to spend extra money, resources and time on a game that already sold and wont sell more as such. Just to spend a lot of work to get it to work on future graphic architectures. Unlike the IHV who have an interest in making it work.

And yes, this is exactly what DX12 is or Vulkan for that matter. Its flawed by design. Great for consoles, disaster on PC.

You are still trying to apply an obscure case to all future cases. Your argument is fundamentally flawed. You are trying to take a game that did a hack job of incorporating an API that wasn't originally designed into the game, and compare it to future games that will be designed with the API fully integrated. You are also assuming that each new release of hardware will have completely different calls for DX12 to handle, which would be the vendors shooting themselves in the foot. I just don't see your argument holding much water at all. Is there more work for the developers? Yes. But will it ultimately provide a better path forward to take better advantages of hardware, especially in the future? Yes.

If you want to make the argument for how beneficial that will be for performance in the end, that is a valid argument. However, we haven't had many good examples of games developed specifically for DX12 yet. And we still won't probably for another year. We have seen some benefits in very specific games for both DX12 and Vulkan. So we know there are benefits. It remains to be seen how that plays out over the next year or 2 now.

I was trying to explain this to our summer intern tester.

Pretend you want a cheeseburger meal.
I can tell my intern: "I want a cheeseburger meal, go get me one!". That's my API call. No matter whether he goes to In-n-Out or McDonalds, I get a cheeseburger meal. We can start to nitpick about the cheeseburger from In-n-Out or McDonalds, like its cheese, or ketchup, or onions, or burger patty(image quality), but it's still a cheeseburger.

That's DX11.

Now with DX12, I have to go buy my own. I can go to either McDonalds or In-n-Out on my own to get the cheeseburger. Now, if I want to, McDonalds or In-n-Out can even let me customize my own cheeseburger! No pickles? No problem! Salad cream instead of ketchup? It's your poison.....

Then, you can walk into a KFC, which doesn't have cheeseburgers.....


Edit: I know it's a terrible analogy, but it was funny when I tried it on my intern. :p

Yes, but that is KFC's fault for not offering cheeseburgers. Because I mean who doesn't want cheeseburgers? And in Shintai's example it doesn't even apply. Because the KFC in his example is Intel, which does offer cheeseburgers, but you didn't know how to ask for them.
 
You are still trying to apply an obscure case to all future cases. Your argument is fundamentally flawed. You are trying to take a game that did a hack job of incorporating an API that wasn't originally designed into the game, and compare it to future games that will be designed with the API fully integrated. You are also assuming that each new release of hardware will have completely different calls for DX12 to handle, which would be the vendors shooting themselves in the foot. I just don't see your argument holding much water at all. Is there more work for the developers? Yes. But will it ultimately provide a better path forward to take better advantages of hardware, especially in the future? Yes.

If you want to make the argument for how beneficial that will be for performance in the end, that is a valid argument. However, we haven't had many good examples of games developed specifically for DX12 yet. And we still won't probably for another year. We have seen some benefits in very specific games for both DX12 and Vulkan. So we know there are benefits. It remains to be seen how that plays out over the next year or 2 now.

You seem to confuse DX12 with a higher layer API like DX11. Its not as easy as that. If it was, we wouldn't have this conversation to begin with. And then DX12 games would work on any IHV as default.
 
You seem to confuse DX12 with a higher layer API like DX11. Its not as easy as that.

Care to give a specific example how I am doing that other than your obscure posts and "1" singular example so far?
 
Back
Top