NTMBK
Lifer
- Nov 14, 2011
- 10,286
- 5,236
- 136
Sorry, to busy reading up on OpenGL, Draw Calls..and reporting posts.
Merry Christmas to you too!
Sorry, to busy reading up on OpenGL, Draw Calls..and reporting posts.
Why are you persisting with this draw call stuff? Why not concentrate on the other parts like asynchronous compute, the new memory model, asymmetric crossfire etc?
We've already seen what the increase in draw calls will do with the Oxide StarSwarm demo. Unless you can show me anything remotely close to that in a gaming sense on another API then it doesn't exist - and no those canned demos of spinning blobs doesn't really count as anything like a game.
That's not really what should happen. One part of mantle allows for 8 cores to work if required, and reducing the draw call CPU penalty should remove the AMD CPU bottleneck.
If it works as planned, CPU's will become near irrelevant to overall performance in many games as the bottleneck shifts to the GPU in almost all cases.
In a nutshell, improved graphics, and more easily. Instead of devs being forced to use "instancing" (which is basically using the same graphic multiple times) because of the DirectX draw call issue, they will be able to use different graphics far more often.
By removing this overhead they actually make it easier for the really good artists to make much better graphics, and the really good coders no longer have to employ so many tricks and hacks to get it to work.
http://www.bit-tech.net/hardware/graphics/2011/03/16/farewell-to-directx/2
Merry Christmas to you too!
/Off topic.
I am from Denmark, we doin not celebrate "Christ-mass", we celebrate "Jul" (Yule), but I fail to see the relevance here?
Besides...it was yesterday...it's over.
Because the draw calls have been used to advance "mantle" even before mantle was announced.
But nice to know that certains aspects of mantle, that now compared to eg. OpenGL dosn't seem all that stellar, now are "offtopic"?
I mean, you use draw calls as an agument?
So, sorry, but I will continue to focus on draw calls.
Because the draw calls have been used to advance "mantle" even before mantle was announced.
But nice to know that certains aspects of mantle, that now compared to eg. OpenGL dosn't seem all that stellar, now are "offtopic"?
Yes the problem is you continue to use one canned benchmark as "proof" of something. Show me those blobs firing at other blobs, acting with some kind of intelligence while maintaining 200 fps. Or even 20 fps.I mean, you use draw calls as an agument?
And you'll be linking that same video as proof of something and still be the only person believing it's anything like the same as what Mantle is capable of.So, sorry, but I will continue to focus on draw calls.
With no official word otherwise, BF4 Mantle should be arriving within 6 days.
No hard data yet so I suggest a harmless guessing game, the 290x vs. 780ti on BF4 Mantle based on HardOCP's apples to apples average frame rate difference between those two cards. This is the dead certain shootout HardOCP will go for first.
My guess, based purely on AMD's CTO saying the 290x will 'ridicule' Nvidia's Titan in BF4 Mantle is a 27% FPS difference, advantage 290x.
Whoever is closest get's a 'bloody well done'.
What is your guess?
My guess is it won't be possible to get an apples to apples comparison.
That depends on what "Ultra" under Mantle is. It could only be possible with say asymmetric crossfire + Kaveri, it could have increased view distances, different methods of lighting which may look better than Ultra on DX, or say a compute based anti-aliasing that is much higher quality but gives a bigger frame-rate hit. So many things can change so that it's not apples to apples in image quality, at which point fps totals become meaningless.
Another thing that could happen (and probably will) is that the Mantle patch removes all of the stalls, huge drops in fps when turning or when there's a lot of explosions on screen. This could be what repi meant with "better gaming experience" - https://twitter.com/repi/status/383012110592376832 but it could also be barely noticeable in the fps totals. A couple of seconds dips over a 2 minute benchmark can't be seen obviously in fps totals but can still have a profound effect on your game experience, especially in multiplayer.
Because the draw calls have been used to advance "mantle" even before mantle was announced.
But nice to know that certains aspects of mantle, that now compared to eg. OpenGL dosn't seem all that stellar, now are "offtopic"?
I mean, you use draw calls as an agument?
So, sorry, but I will continue to focus on draw calls.
Lots to consider there, yet I can only assume HardOCP will follow it's usual format and do an apples to apples comparison based on equalizing settings as best they can. I doubt there will be a major review site that doesn't attempt to quantify the difference caveated though it may be.
That depends on what "Ultra" under Mantle is. It could only be possible with say asymmetric crossfire + Kaveri, it could have increased view distances, different methods of lighting which may look better than Ultra on DX, or say a compute based anti-aliasing that is much higher quality but gives a bigger frame-rate hit. So many things can change so that it's not apples to apples in image quality, at which point fps totals become meaningless.
Another thing that could happen (and probably will) is that the Mantle patch removes all of the stalls, huge drops in fps when turning or when there's a lot of explosions on screen. This could be what repi meant with "better gaming experience" - https://twitter.com/repi/status/383012110592376832 but it could also be barely noticeable in the fps totals. A couple of seconds dips over a 2 minute benchmark can't be seen obviously in fps totals but can still have a profound effect on your game experience, especially in multiplayer.
Yes and those metrics are image quality and playability at those levels of quality.
Example - If the game is already making decent use of multi-core CPU's like BF4 appears to be, you can still make even better (full) use of them by increasing draws etc. That might lower the fps totals slightly but make the game look much better. How do you quantify the experience? It's like with Crossfire/Sli vs single-card - just because the fps total is higher doesn't mean it's a better experience. It rarely is.
And where does it end? One card or four? I'm fairly sure AMD will get near-perfect scaling with quadfire making "ridicule" a guarantee. Is that shifting the goalposts as well?
At least for me Battlefield 4 can do 110-177 fps and is still GPU limited. It shows no signs so far of being draw call limited. Now what may well be happening is that the fidelity of the image was decreased to fit into DX limits, but it ends up a decent looking game. A direct conversion to Mantle doesn't look like it would benefit much, the game is completely limited by GPU performance. Admittedly if you push even more GPU performance (a pair of 780ti's at 1080p) then you might very well hit the CPU and draw call limits, but its a game already achieving 110 fps in an intense moment not withstanding GPU limits.
BF4 doesn't look like a good game for draw call overhead reduction to make an appreciable difference in fps from my own testing on 64 player games. I can't predict the future or the impact but BF4 isn't one of those games I consider has a problem with draw calls and hence it isn't likely to be a good poster child for what is possible.