Possible and practical are 2 different ways. But in terms of DX its very simple, no hardware, no support.
And as you quote:
And as you quote:
However both approaches add performance overhead, and as such usage of conservative rasterization in real time graphics has been pretty limited so far
Possible and practical are 2 different ways. But in terms of DX its very simple, no hardware, no support.
So what does this mean in regards to Nvidias "Asynchronous Compute support" ?
Possible =/ practical
For all intent and purposes, Nvidia don't support Async compute.
pretty big visual difference between HBAO+ and VXAO here
although part of the reason why is that it looks like HBAO+ isn't being applied to the vegetation
Yeah, VXAO seems like a very good AO solution.
And, in all honesty, the only thing that's showing as genuinely slower on those DX 11 vs 12 charts seems to be fury.
I would not be hugely surprised if these sorts of charts end up being what most programmed DX12 games to end up looking like.
Yes the API lets you get non trivial advantages for specific GPU architectures if you really put big effort into optimising at a low level for that architecture.
Given the amount of 'effort' seemingly put into PC ports, would anyone expect them to do that? For, the ~half dozen architectures there will be the PC space?
(For GCN 1.1 yes, but that architecture is gone in 3-6 months time.).
Especially given that all that effort is wasted/even counter productive in 2/3 years time when all the cards on sale are using different architectures.
I imagine that the CPU benefits should be much safer to get.
PS - Serious question. Has anyone checked to see what all these DX12 things do on Intel iGPU's? That would be quite an interesting sort of data point. And, yes, given trends we do need to take those reasonably seriously too.
Is this that bad?
The charts overleaf looks like a <5% loss for most of the situations/GPUs. Not ideal but I'm unsure if that counts as hugely significant Probably balanced by the gains in terms of weaker CPUs.
GCN 1.2 (380/Fury) does look to need more work for some reason.
Certainly if DX12 can be implemented in a manner which only drops ~5% for new architectures, with options for heavy optimisation for some specific ones, it'll be fine.
It should be obvious to even the trolliest of the trolls, that when you consider DX11 970 is handily beating DX12 970, that this DX12 implementation is horrendously broken... And this is coming form an NVidia owner.
Considering it's an API that should do the exact opposite of what it's doing, yes it's that bad. What the heck is the point of DX12 otherwise?
I suppose if you were playing this game in DX11 and you just had way too much performance that it bothered you, DX12 would be useful. I haven't heard of any such complaints though.
This was my concern with DX12 as well, I wasn't convinced devs would put in the effort required to really take advantage of it. Even still, i'm surprised it's THIS bad. I did not expect significantly worse performance, I just didn't expect some of the "possible" features some people were excited about. Like being able to use any combination of cards for SLI/CF, or being able to double up on your usable vram when running SLI/CF
I know DICE got a lot of flak for their implementation of Mantle where a lot of people had issues, but in hindsight it's probably the best implementation of a "patched in" API we have yet. It may have had some bugs, but it actually did what it set out to do.
I think you wanted to mention Unity. Oh yes, Unity had some issues.I reckon Assassin's Creed Black Flag didn't get any major performance issues being a Gameworks title.
Considering it's an API that should do the exact opposite of what it's doing, yes it's that bad. What the heck is the point of DX12 otherwise?
I suppose if you were playing this game in DX11 and you just had way too much performance that it bothered you, DX12 would be useful. I haven't heard of any such complaints though.
Leave aside GCN 1.2 which seems to need patching or something but still isn't collapsing.
Otherwise, as far as I can see - and I might be wrong - you're dropping ~5% at worst, and people with lower powered CPUs are seeing quite big frame rate increases?
You'd obviously prefer it to be +10% than -5% or something - and imagine once they're really used it to they might manage it - but it's OK. Feels like what you'd expect from DX12 done plausibly competently but without specific architectural optimisations.
(Which is likely how most PC gamers will end up experiencing it.).
The gears of war remake was much more troubled.
Mind you, I would still like to know how it goes on Intel's IGP
It shouldn't have a drop at all. It's suppose to be more efficient so at worst, meaning if you had no CPU bottleneck at all at any point in the game, you should get the same performance. When you're dropping frames almost across the board, yes, it's that bad. In this game, DX12 isn't doing what it's supposed to do. It's doing the opposite. I know I sound like a broken record, but it's also the second time you made a nearly identical statement.
It's like going into a job expecting a promotion and raise. You get the promotion but a pay cut instead. Not that bad is it?
Synchronization between threads is a bitch,you see it in games all the time,the more threads the more synchronization the more you drop.It shouldn't have a drop at all.
...
The point of DX12 is CPU overhead reduction, specially for Xbox One. ..blahblah blah
stopped reading right there as it is obvious you DON'T know the point of DX12.. please do yourself and everyone else a favor and actually read up on DirectX 12 before commenting
https://blogs.msdn.microsoft.com/directx/2014/03/20/directx-12/
https://msdn.microsoft.com/en-us/library/windows/desktop/dn903821(v=vs.85).aspx
https://www.youtube.com/channel/UCiaX2B8XiXR70jaN7NK-FpA
What makes Direct3D 12 better? First and foremost, it provides a lower level of hardware abstraction than ever before, allowing games to significantly improve multithread scaling and CPU utilization.
Nope, you can't mine well in those chips compared to AMD cards due the tech used on them.Not practical. Just like mining on NVidia cards.
Nope, you can't mine well in those chips compared to AMD cards due the tech used on them.
Why shouldn't it have a drop at all? It's a new API, if you aren't using it correctly you very well will get a drop in performance compared to another API you use correctly. How you write efficient quality code for the new API may very well be different than what you did for the old API. Just because the code works doesn't mean it works well.
Hopefully GDC this week will get some more progress in the right direction.
I am interested to see what we see from Frostbite and DX12.
It should be obvious to even the trolliest of the trolls, that when you consider DX11 970 is handily beating DX12 970, that this DX12 implementation is horrendously broken... And this is coming form an NVidia owner.
Is this that bad?
The charts overleaf looks like a <5% loss for most of the situations/GPUs. Not ideal but I'm unsure if that counts as hugely significant Probably balanced by the gains in terms of weaker CPUs.
GCN 1.2 (380/Fury) does look to need more work for some reason.
Certainly if DX12 can be implemented in a manner which only drops ~5% for new architectures, with options for heavy optimisation for some specific ones, it'll be fine.
DX12 is purely for performance enhancement. If it's not doing that then it's not done properly, either on the software or hardware side (or both, possibly).
As someone else pointed out, AMD and it's partners have more experience with prior work on Mantle. Maybe it'll just take a bit of time for everyone else to work it out.