96Firebird
Diamond Member
- Nov 8, 2010
- 5,712
- 316
- 126
By the time we need these features, if we ever do, we will all be on to the next gen cards.
Smells like tessellation all over again...
We all know how well that worked out for AMD. :awe:
As expected. Features NV suck at = not important. So predictable.
I guess when Pascal or Volta is good at it, Async Compute will be important again.
You're probably right. AMD's timing is always way off. So, in this case, perhaps by the time Async becomes important (if ever) All offered Nvidia hardware will run it better than AMD will. This is a repetitive story line. A stale story line.
Firebird is exactly right. Just like tesselation.
AMD introduced Tru-Form {tesselation}. It was cool, but nobody used it. But by the time it became present in games, Nvidia handed AMD it's own arse.
So please Silver, consider history for a moment and stop acting like AMD is a brand new company and you have no knowledge of past let downs.
Every new thing AMD introduces is like the second coming. I mean WITAF?
What I really meant to say was, that Maxwell's architecture is just more efficient. I should not have used the word pipeline.. Maxwell GPUs have less transistors than comparable GCN GPUs, but still manages to outperform them at lower TDPs.
Maxwell architecture also scales better than GCN 1.2.
Mate, a few weeks ago Async Compute was the second coming. Devs are hyping it up, sites covered it massively... and now... when disappointment wells up as their hardware which they were led to believe was great at it falls on its face, people just can't deal with it. So what do they do? They tear it down.
When Tru-Form was around, did developers hype that up?
Just look at the games coming out in DX12, many of them, their own studios/devs are hyping up async compute.
The only WTF-worthy here is why is NV still silent, where's their PR statement?
Mate, a few weeks ago Async Compute was the second coming. Devs are hyping it up, sites covered it massively... and now... when disappointment wells up as their hardware which they were led to believe was great at it falls on its face, people just can't deal with it. So what do they do? They tear it down.
When Tru-Form was around, did developers hype that up?
Just look at the games coming out in DX12, many of them, their own studios/devs are hyping up async compute.
The only WTF-worthy here is why is NV still silent, where's their PR statement?
That's all it is right now. Hype. How many freight cars are on the AMD hype train? You'd be sitting at the crossing for a long while.
How about this, until Async on AMD shows practical application advantages over it's competition on a regular basis and is widely accepted as the de-facto way to do things throughout DX12, try not to add to many more cars to the hype train. There's graffiti on the train and it is unsightly.
I'm telling you. You do not learn from history. If it IS as important as you claim it is, Nvidia will just turn around and support it stronger, and you'll regret all this promotion and cheerleading. How many times does this have to happen to you to realize that this is the norm? When will you stop hyping?
Go look at past threads when DX11 was announced. Hell, Dirt 2 was delayed to add DX11 features, including tessellation. A developer wouldn't delay a game to add a feature they weren't actively promoting... Not to mention the countless demos and Q&A's about DX11 and tessellation that ATi/AMD promoted.
Looking back at these threads, you can also see the hope that DX11 brought to gamers. People were claiming DX11 would save the industry, and it was everything DX10 should have been. What happened to that? Why is DX11 now not enough, and will DX12 really be any different?
The similarities are astonishing, but doesn't give much hope.
At least you've noticed the hype for Async Compute. There was zero dev-generated hype of Tru-Form, so you history lesson already has brought light to the different scenario.
NV is better at tessellation, great. It's not a concern of mine, I can set whatever tessellation factor I want to avoid abuse. I really enjoyed HairWorks in Witcher 3, at x16 factor for months before it was made official in a patch. Now even NV users can appreciate the performance gains.
Btw, since I was one of the few who spoke out about Ryan Smith's Async article, about the misleading table, or how Maxwell actually has 1 engine with 32 compute queues and not 32 engines. When I questioned whether Maxwell can even do Async Compute properly, many of you attacked me personally in those threads. You now call me a cheer leader.
Ad Hominem is the last refuge of a scoundrel, as they say.
Looking back at these threads, you can also see the hope that DX11 brought to gamers. People were claiming DX11 would save the industry, and it was everything DX10 should have been. What happened to that? Why is DX11 now not enough, and will DX12 really be any different?
The similarities are astonishing, but doesn't give much hope.
Here's the difference.
http://forums.anandtech.com/showpost.php?p=37678573&postcount=52
Read it, it comes from the devs.
None of the links work, and I can't access YouTube at work.
Edit - Looks like you just copied and pasted, meaning the links are shorthanded and didn't copy over.
That was 5 or 6 years ago. The problem with DX11 is it's serial nature.
What it IS, is a way to parallelize compute functions instead of serialize them. Look in the Ashes thread. You've been shown that Ashes is the exception and no other game will ever use that much content relying on Async. It's only one game from a dev who supports AMD. Of course it's going to shine like that god beam.
Whatever the case. To Async or not to Async is up to the devs, and if it looks like it's going to become widely used, then AMD won't be the only ones who do it well.
And then what will Async be to you? Second coming? Or just another tesselation?
Go look at past threads when DX11 was announced. Hell, Dirt 2 was delayed to add DX11 features, including tessellation. A developer wouldn't delay a game to add a feature they weren't actively promoting... Not to mention the countless demos and Q&A's about DX11 and tessellation that ATi/AMD promoted.
Looking back at these threads, you can also see the hope that DX11 brought to gamers. People were claiming DX11 would save the industry, and it was everything DX10 should have been. What happened to that? Why is DX11 now not enough, and will DX12 really be any different?
The similarities are astonishing, but doesn't give much hope.
Why do you think this is? Because Mantle and DX12 are essentially the same? Or did AMD somehow predict what Microsoft would do with DX12 and Nvidia didn't.
Or more utilized under DX11.
To reasonable of an assessment. Clearly there is a secret evil hidden agenda!I think both companies had a lot of say and input about dx12 with MS. The manufacturers just prioritize different things. Like nvidia probably felt it was better to focus on one area and amd focused elsewhere. That lands us here today where there are different things each excels at over the competition.
That's all it is right now. Hype. How many freight cars are on the AMD hype train? You'd be sitting at the crossing for a long while.
How about this, until Async on AMD shows practical application advantages over it's competition on a regular basis and is widely accepted as the de-facto way to do things throughout DX12, try not to add to many more cars to the hype train. There's graffiti on the train and it is unsightly.
I'm telling you. You do not learn from history. If it IS as important as you claim it is, Nvidia will just turn around and support it stronger, and you'll regret all this promotion and cheerleading. How many times does this have to happen to you to realize that this is the norm? When will you stop hyping?
How about this, until Async on AMD shows practical application advantages over it's competition on a regular basis and is widely accepted as the de-facto way to do things throughout DX12
OK I concede your point. But the performance gain is really dependent on the architecture of the GPU, and whether it has dedicated hardware resources or not..
As I've been saying, GCN obviously has a lot to gain from using asynchronous compute. But whether Maxwell can, is another matter entirely.
Personally I think Maxwell does support asynchronous compute, otherwise the beyond3d test wouldn't show the 31 queue dispatch:
WTF I did just read... I mean, you don't need to be a genius to understand that amd pushing for AC is the first step that needs to be done for AC to be "widely accepted as the de-facto way to do things throughout DX12".
Based on what I've gathered in the b3d forums, there is an issue with terminology here. Both Maxwell and Kepler can do asynchronous dispatch of compute threads (what most, including Nvidia, would call async compute).Then explain why the simultaneous compute and graphics graph, is about equal to the individual compute and graphics times added together?
What ad Hominem? I'm just surprised you do not learn to stop supporting the hype train after so many times.
I like talking about new tech, but give it the credit it deserves. Not this God Beam of sunlight stabbing a hole through the
dark dismal sky shining just the right amount of light to stave off the evil forces taking over the dark gaming industry.
It ain't.
What it IS, is a way to parallelize compute functions instead of serialize them. Look in the Ashes thread. You've been shown that Ashes is the exception and no other game will ever use that much content relying on Async. It's only one game from a dev who supports AMD. Of course it's going to shine like that god beam.
Whatever the case. To Async or not to Async is up to the devs, and if it looks like it's going to become widely used, then AMD won't be the only ones who do it well.
And then what will Async be to you? Second coming? Or just another tesselation?