Upcoming/possible games using Mantle

[Skurge]

I don't see domination or CPU limited scenarios in this BETA game.

 

[Headfoot]

Yes because Star Citizen is not AMD evolved or Nvidia sponsored.
I would say again u should watch that interview as i stated above it will be very helpful for u.
Peace.

I don't care what one guy says in a multi-billion dollar company, especially when that one guy isn't the CEO

 

[DiogoDX]

In short he is saying Nvidia will totally dominate if game uses Multi threading and that we seen in BF4 and Battlefield Hardline
https://translate.google.com/translate?sl=auto&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=http%3A%2F%2Fgamegpu.ru%2Faction-%2F-fps-%2F-tps%2Fbattlefield-hardline-beta-test-gpu.html&edit-text=&act=url]
Just look at cpu usage by of both companies than u will know the difference.

A 23 seconds benchmark is not representative of actual multiplayer gameplay.:thumbsdown:

Much better test: http://www.pcgameshardware.de/Battl...Battlefield-Hardline-Beta-Benchmarks-1125079/

780TI Phantom with +20% overclock than stock 780TI

 

[3DVagabond]

I always have to explain this about two or three times I swear.. :\

The performance boost derived from DX11 multithreading only manifests itself in CPU limited scenarios. Whether a system is CPU limited or not depends on several factors such as IQ settings, resolution, and how powerful the GPU is.

At 1080p with 4x MSAA the GTX 680 is more GPU limited, which is why the performance gain is only 6%. But when you look at a more powerful card like the GTX 780, the performance gain over the 7970 GE is 46%.

The DX11 multithreading has a bigger impact on the GTX 780 because it's the most CPU limited card in that assortment..

Well, if we can discuss this in an upcoming Mantle games thread I suppose we could put this here as well?

This is Hardware.fr showing results at lower resolution and how it effects relative performance.

Here is the relative performance @ 4K using the same settings.

These are the biggest variation games (excluding Batman AO for obvious reasons). I would say that it is more than a driver overhead advantage that nVidia is showing at lower resolutions. More than likely nVidia having issues with their drivers @ 4K and the cards not being pushed as hard causing less throttling of the 780 ti SLI setup is influencing results as well. Never the less, it does show consistently higher performance at lower resolution and that could be attributed to lower driver overhead in general, which is what Hardware.fr is saying.

What, if any games here are using DX11mt?

For those who want to examine the entire review, LINK TO SOURCE

 

[3DVagabond]

In short he is saying Nvidia will totally dominate if game uses Multi threading and that we seen in BF4 and Battlefield Hardline
https://translate.google.com/translate?sl=auto&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=http%3A%2F%2Fgamegpu.ru%2Faction-%2F-fps-%2F-tps%2Fbattlefield-hardline-beta-test-gpu.html&edit-text=&act=url]
Just look at cpu usage by of both companies than u will know the difference.

OMG! I wonder why AMD wouldn't optimize their drivers running DX11 in Battlefield 4?

You might just want to link to the part you are trying to use to demonstrate your point. Rather than make people read the whole article and hope they find what you want them to.

I don't care what one guy says in a multi-billion dollar company, especially when that one guy isn't the CEO

Add to that He has it wrong. Huddy said that the devs aren't allowed to share with AMD anything that will allow them to optimize their hardware to nVidia's proprietary features. As far as anything else it likely would vary from one contract to the next.

 

[rgallant]

And Mantle runs only on certain AMD hardware. That is less than <20% of the whole market.
The question is: Can you afford to ignore the non Mantle market? I hope so because otherwise praising a solution to a non existing problem sounds more like marketing.

What? Only paid and child companies from the console manufactures are only selling their games on consoles. Every other company will release a PC version. Even EA is backpedalling with their sports series like Fifa.

is that like gsync with user add in cards maybe 200 total monitors vs/out of 600 million but yet is the greatest gameing add on ever only $800+ when asus monitor gets released.can't even do the math on that.
forgot to add -and only runs on nv cards

 

[desprado]

OMG! I wonder why AMD wouldn't optimize their drivers running DX11 in Battlefield 4?

You might just want to link to the part you are trying to use to demonstrate your point. Rather than make people read the whole article and hope they find what you want them to.


Add to that He has it wrong. Huddy said that the devs aren't allowed to share with AMD anything that will allow them to optimize their hardware to nVidia's proprietary features. As far as anything else it likely would vary from one contract to the next.

Its simple because Nvidia driver Suppport Multi threading and know with some new updates to there drivers Multi threading is more efficient.

 

[desprado]

A 23 seconds benchmark is not representative of actual multiplayer gameplay.:thumbsdown:

Much better test: http://www.pcgameshardware.de/Battl...Battlefield-Hardline-Beta-Benchmarks-1125079/

780TI Phantom with +20% overclock than stock 780TI

What about MSI R9 290X GE lol with Mantle enable and it should beat nvidia DX11.

I know many user wont agree but Mantle is to just help AMD gpu because it lack of Multi threading support in DX11.

 

[Deders]

I know many user wont agree but Mantle is to just help AMD gpu because it lack of Multi threading support in DX11.

And AMD cpu's. AMD have re-written the rules to maker the best use of their hardware all round.

 

[desprado]

And AMD cpu's.

It does not even exist in competition so leave it.

 

[desprado]

This is CPU usage in BF Hardline by using Nvidia GTX 780 TI on Dx11

This is CPU usage in BF Hardline by using R9 290X using Dx11

 

[sontin]

It is much more fun to compare nVidia's DX11 driver in non Mantle games like Grid:Autosport:
http://www.pcgameshardware.de/Grid-...karten-Benchmarks-CPU-Kernskalierung-1126276/

A custom GTX780TI is 50% faster than a custom 290X.
But yeah we know that DX11 multithreading (!= DCL) is "complete broken".

is that like gsync with user add in cards maybe 200 total monitors vs/out of 600 million but yet is the greatest gameing add on ever only $800+ when asus monitor gets released.can't even do the math on that.
forgot to add -and only runs on nv cards

Lolwut? G-Sync works in every game. Mantle in three.
Not even the point that G-Sync is a hardware solution to real problems.
When your "next gen" API gets beaten by DX11 then there is nothing else to discuss.

 

[Carfax83]

7970 has memory clocked at 1375. The 7970Ghz has memory clocked at 1500. The 7970 has memory clocked 92% as fast as a 7970GHz.

Consequently, it scores 90% of what a 7970Ghz does. Fudging it a little to account for the slight core clockspeed difference and benchmark variance, you can clearly see the biggest bottleneck in the Tahiti based GPUs on that graph isn't whatever CPU they used, but rather the GPU memory clock speed. Well you can't, obviously, because despite having that chart saved in your "AMD FUD" bookmarks group and posting it every chance you get it is pretty clear that you have never looked at it critically.

Right, so I'm to believe that memory bandwidth plays such a great role in this benchmark, that the GTX 780 which has the SAME effective bandwidth as the HD 7970 GHz has a whopping 46% lead, when it's on average about 20% faster than the HD 7970 GHz

And the GTX 770 with less bandwidth than the 7970 GHz and lower compute performance still has a significant performance advantage over the latter?

Are there other factors that play a role in the GTX 780s dominance in that benchmark other than DX11 MT? Sure...

But because of your zeal you cannot admit that the lack of DX11 MT is also hurting AMD big time in that benchmark..

 

[Carfax83]

7970GE should not be that much faster than 7970 using that logic. The performance difference between the 680 and the 670 is the same as the 7970GE and the 7970.

Of course there are other factors in play here rather than just DX11 MT.. But DX11 MT plays a big role, because it allows the GPU to approach maximum performance..

The GTX 780's lead is almost 50%, which is huge and way over the average 20% advantage it usually has over the HD 7970 GHz.. And then we have the GTX 680 with significantly less bandwidth and compute performance than the HD 7970 GHz, yet still managing to beat it in a decisive victory.

It's easy to claim the 780 is faster cause of MT, but until you can compare it with a 290 in that game the gap is similar. I don't think you can make that conclusion.

It's too bad that Anandtech no longer uses Civ 5 benchmark. I would have loved to see them bench the R9 290 series..

 

[3DVagabond]

Never mind. I don't know enough about the game to really comment.

 

[Carfax83]

These are the biggest variation games (excluding Batman AO for obvious reasons). I would say that it is more than a driver overhead advantage that nVidia is showing at lower resolutions. More than likely nVidia having issues with their drivers @ 4K and the cards not being pushed as hard causing less throttling of the 780 ti SLI setup is influencing results as well. Never the less, it does show consistently higher performance at lower resolution and that could be attributed to lower driver overhead in general, which is what Hardware.fr is saying.

Not only lower overhead, but more effective use of CPU cores via driver optimizations.

What, if any games here are using DX11mt?

None of them are to my knowledge. But a game doesn't have to be using DX11 MT for NVidia to have an edge due to their aggressive DX11 optimizations and leveraging of the CPU.

 

[3DVagabond]

Not only lower overhead, but more effective use of CPU cores via driver optimizations.



None of them are to my knowledge. But a game doesn't have to be using DX11 MT for NVidia to have an edge due to their aggressive DX11 optimizations and leveraging of the CPU.

Well why are attributing it to DX11 mt then if these games don't use it?

 

[sontin]

Never mind. I don't know enough about the game to really comment.

I'm talking about Grid:Autosport and not Grid 2.

Well why are attributing it to DX11 mt then if these games don't use it?

DX11mt is nothing else than a multithreading engine. Nearly every new advanced engine can use up to 8 threads.
Star Swarm is using every available thread to improve the performance - even without Deferred Context.

 

[desprado]

Well why are attributing it to DX11 mt then if these games don't use it?

Assassins Creed 4 use MT,Star Swarm and even Frostbite engine 3 use multi rendering.

 

[3DVagabond]

I'm talking about Grid:Autosport and not Grid 2.



DX11mt is nothing else than a multithreading engine. Nearly every new advanced engine can use up to 8 threads.
Star Swarm is using every available thread to improve the performance - even without Deferred Context.

What part of "never mind I don't know the game well enough to comment", were you responding to?

So, are these games using it or not? The answer seems to change depending on whether it helps make someone's point or not.

 

[desprado]

What part of "never mind I don't know the game well enough to comment", were you responding to?

So, are these games using it or not? The answer seems to change depending on whether it helps make someone's point or not.

Bro u can look at the benchmark that i posted above u will get ur answer all most all upcoming AAA game will be using multi rendering.

 

[3DVagabond]

Bro u can look at the benchmark that i posted above u will get ur answer all most all upcoming AAA game will be using multi rendering.

3 people have responded now and not one has answered my question. Is it too hard for people to just say they don't know, or they aren't sure, or just not say anything? Or is it just too important to derail a Mantle thread.

 

[AtenRa]

This is CPU usage in BF Hardline by using Nvidia GTX 780 TI on Dx11

This is CPU usage in BF Hardline by using R9 290X using Dx11

You do see that those graphs are using GTX 780Ti SLI and R9 295x2 ??
This has nothing to do with DX-11mt but more about Driver SLI and CF optimization at the BETA stage.

Same review, single R9 290X produce almost the same fps as Dual GPU R9 295x2 from above.

 

[desprado]

3 people have responded now and not one has answered my question. Is it too hard for people to just say they don't know, or they aren't sure, or just not say anything? Or is it just too important to derail a Mantle thread.

Yes there are few games that Multi rendering like Assassins Creed 4,FrontBite 3 games engine,Crysis 3.

 

[desprado]

You do see that those graphs are using GTX 780Ti SLI and R9 295x2 ??
This has nothing to do with DX-11mt but more about Driver SLI and CF optimization at the BETA stage.

Same review, single R9 290X produce almost the same fps as Dual GPU R9 295x2 from above.

u did not see single GTX 780 TI?It has shown all gpu benchmark.