The AMD Mantle Thread

[ASM-coder]

A lot of people on the battlelog forums stating that the Mantle renderer looks different with reduced draw distances

Someone on this thread(Paul???) mentioned a few days ago that draw distances seemed to have decreased with the new BF4 patch. This was not Mantle related.

 

[Cloudfire777]

All of these benchmarks released the latest 24 hours have made a complete mess out of the whole Mantle launch.

The results range from +50% down to +2%. Some consistency with enforcing similar testing protocols on all reviewer sites would have been the right move. Then allow reviewers to freely explore other systems later once the first data have settled and sunk in.

Right now I`m very confused where we stand

 

[SiliconWars]

http://battlelog.battlefield.com/bf4/forum/threadview/2979150493821981180/1/

A lot of people on the battlelog forums stating that the Mantle renderer looks different with reduced draw distances or way more fog than DX11. Maybe the beta nature of the driver.

Looks like a lighting issue but everything is being rendered.

It is in fact not being used in all FB3 games. NFS Rivals specifically is not getting it despite using FB3. To my knowledge, Dragon Age: I is being released this year and the developers have not confirmed Mantle. PvZ Zombie Warfare is also unconfirmed for Mantle by the developers, although the bigger question is why would that game need it in the first place? Not a demanding game, more oriented to the super casual crowd. We'll see, I guess, if the developers confirm Mantle or not.

According to http://www.slideshare.net/DICEStudio/mantle-for-developers slides 37 and 38, all those games will have Mantle.

 

[DarkKnightDude]

I remember the Bioware devs saying that DA3 and the new Mass Effect will use it for sure.

I did notice that the game seems a tad brighter in Mantle mode, maybe just me, but looks about the same as DirectX

 

[SiliconWars]

All of these benchmarks released the latest 24 hours have made a complete mess out of the whole Mantle launch.

The results range from +50% down to +2%. Some consistency with enforcing similar testing protocols on all reviewer sites would have been the right move. Then allow reviewers to freely explore other systems later once the first data have settled and sunk in.

Right now I`m very confused where we stand

There's really not much to be confused about. In GPU limited situations the gains will be around 10%. CPU limited scenarios (eg BF4 in multiplayer) can potentially have massive gains but are generally kicking around the 20%-30% mark because you're not ALWAYS going to be CPU limited in multi-player either - in fact you might spend a lot of time GPU limited running from spot to spot. On the flipside you might end up in a protracted battle with 50 people and being completely CPU bound - in which case Mantle will shine.

That's why the results are all over the place and always will be regarding multi-player.

 

[ASM-coder]

PvZ Zombie Warfare is also unconfirmed for Mantle by the developers, although the bigger question is why would that game need it in the first place? Not a demanding game, more oriented to the super casual crowd.

As I said before, a Frostbite slide from a Johan Andresson presentation states clearly that
PvZ Garden Warfare would support Mantle, and the slide stated that they were targeting the APU market.
Why post on discussion boards, when you have already made up your mind, and are not open to anything that contradicts your beliefs?

 

[Grooveriding]

Gamegpu.ru Mantle results

http://gamegpu.ru/action-/-fps-/-tps/battlefield-4-mantle-vs-directx-11-test-gpu.html

DX11

Mantle

 

[SiliconWars]

http://battlelog.battlefield.com/bf4/forum/threadview/2979150493821981180/1/

A lot of people on the battlelog forums stating that the Mantle renderer looks different with reduced draw distances or way more fog than DX11. Maybe the beta nature of the driver.

Repi just updated, says it's a bug - http://battlelog.battlefield.com/bf4/forum/threadview/2979150493821981180/3/

2014-02-02 14:31
Thanks for reporting, it is a bug on our side that came in as part of a late change, know what it is and we'll fix it. Doesn't affect performance.

It's also kinda weird that in the screenshots you linked, Mantle seems to have a higher level of shadowing (under roofs are plain to see).

 

[caswow]

http://battlelog.battlefield.com/bf4/forum/threadview/2979150493821981180/1/

A lot of people on the battlelog forums stating that the Mantle renderer looks different with reduced draw distances or way more fog than DX11. Maybe the beta nature of the driver.

Its a bug and they know about it. will be fixed witouth perf. loss. page 3 of your posted thread.

 

[blackened23]

Its a bug and they know about it. will be fixed witouth perf. loss. page 3 of your posted thread.

There was no page 3 at the time I posted the link, so repi apparently just posted that recently. It appears that AMD's extensive social media and forum marketing team did send tweets repi's way to inform him of this very recently, or perhaps he actually saw it here and AMD sent him this way, but he just posted it recently. Again, no page 3 at the time I posted the link. I saw it mentioned at HardOCP and linked it here being curious to see if others experienced the same thing. It's an alpha driver for Mantle after all, so I wasn't sure if it was that or game related.

Thanks for the update though. Apparently the Mantle renderer is rendering things differently, but it is a bug. Which will be fixed. Thanks for keeping us abreast of this news, hopefully it is fixed soon so that those that are able to get the alpha AMD driver working, can get the same graphical level as DX11.

 

[caswow]

There was no page 3 at the time I posted the link, so repi apparently just posted that recently. It appears that AMD's extensive social media forum/marketing team did send tweets repi's way to inform him of this, but he just posted it recently.

Thanks for the update though. Apparently the Mantle renderer is rendering things differently, but it is a bug. Which will be fixed. Thanks for keeping us abreast of this news.

no no i just replied to your post so everybody knows what dice had to say about this

 

[redzo]

r9 280x(7970) mantle performance is by far the most interesting. It hits the sweet spot in terms of price. Die hard BF4 fans will buy this and mantle's gains could completely dethrone the gtx770 alternative purchase. The 280x is also way cheaper than a 290. ... but, strange enough, people seem to be concerned more about low end(kaveri) and high end(290/x).

 

[blackened23]

As I said before, a Frostbite slide from a Johan Andresson presentation states clearly that

I didn't see the PvZ slide. However, the point stands, not all FB3 games are getting Mantle treatment - NFS Rivals being case in point. Hilariously enough, that game has a hard coded 30 fps limit still to this day which is tied to the physics engine. So that brings into question, of the other games being released on FB3, did the developers confirm Mantle or not? I don't know. Right now Thief and BF4 are the known games, PvZ apparently (although this is mostly for low end APU users apparently, as you stated), with other games being unknown. Star Citizen is known, but won't be a complete game with all modules in 2015. We'll see I guess. Perhaps you can take a look into Dragon Age specifically and ask the devs on the bioware forum and keep us up to date on that. Product page has no mention or confirmation of Mantle, I think AMD did make a bucket list of all FB3 games in one of their slides but not all of them are using Mantle simply because they're FB3. Mantle still requires developer time and intervention regardless of using FB3 or not.

 

[desprado]

Looks like a lighting issue but everything is being rendered.

According to http://www.slideshare.net/DICEStudio/mantle-for-developers slides 37 and 38, all those games will have Mantle.

Yes this is true but dont expect Mantle on NSF as it will be 30fps limited from now on.

 

[SiliconWars]

There was no page 3 at the time I posted the link, so repi apparently just posted that recently. It appears that AMD's extensive social media and forum marketing team did send tweets repi's way to inform him of this very recently, or perhaps he actually saw it here and AMD sent him this way, but he just posted it recently. Again, no page 3 at the time I posted the link. I saw it mentioned at HardOCP and linked it here being curious to see if others experienced the same thing. It's an alpha driver for Mantle after all, so I wasn't sure if it was that or game related.

Thanks for the update though. Apparently the Mantle renderer is rendering things differently, but it is a bug. Which will be fixed. Thanks for keeping us abreast of this news, hopefully it is fixed soon so that those that are able to get the alpha AMD driver working, can get the same graphical level as DX11.

No problem. If Nvidia's extensive social media/forum marketing team gives you any more pointers on what FB3 games won't be getting Mantle optimizations, you be sure to let the rest of us know.

 

[SiliconWars]

Yes this is true but dont expect Mantle on NSF as it will be 30fps limited from now on.

What's stopping them from doing a smoother Mantle version though - like we've seen in numerous screenshots with BF4?

 

[brandon888]

ahahah 2 fps gain for 7950 yeah but it's true not even 20% performance gain .... dissapointed ..

 

[ASM-coder]

It appears that AMD's extensive social media and forum marketing team did send tweets repi's way to inform him of this very recently, or perhaps he actually saw it here and AMD sent him this way, but he just posted it recently.

Do you realize how you sound? This is still in Beta, Johan explicitly stated on the roll-out:
"If you try out Mantle in BF4 we would really like to hear your feedback about it! If you have any feedback about performance or your experience in general with Mantle in Battlefield 4, please post on the Battlelog forums or in the comments section below."


Why would this have anything to do with AMD, and why would anyone need to tweet him to look at that site? He directed people to post there.
And if you read through that thread you posted, you get a good taste of a real "marketing team" in action.

 

[Nintendesert]

http://battlelog.battlefield.com/bf4/forum/threadview/2979150493821981180/1/

A lot of people on the battlelog forums stating that the Mantle renderer looks different with reduced draw distances or way more fog than DX11. Maybe the beta nature of the driver.



It is in fact not being used in all FB3 games. NFS Rivals specifically is not getting it despite using FB3. To my knowledge, Dragon Age: I is being released this year and the developers have not confirmed Mantle. PvZ Zombie Warfare is also unconfirmed for Mantle by the developers, although the bigger question is why would that game need it in the first place? Not a demanding game, more oriented to the super casual crowd. We'll see, I guess, if the developers confirm Mantle or not.

Erm.... if Mantle gives better performance on the low end and on APUs that's exactly why a game geared towards the casual crowd would be using it. :hmm:

 

[IEC]

Erm.... if Mantle gives better performance on the low end and on APUs that's exactly why a game geared towards the casual crowd would be using it. :hmm:

+1. Mantle sounds like a great feature for the gamer who doesn't upgrade his CPU very often. Which is the average mainstream gamer.

 

[biostud]

ahahah 2 fps gain for 7950 yeah but it's true not even 20% performance gain .... dissapointed ..

uhm, the driver release notes clearly states that only GCN1.1 cards are optimized ATM (R9 290(X) and R9 260) while 7xxx cards will be optimized with a later release.

For me personally I expect quite an increase once drivers for CF and 7xxx cards are optimized, as I run with an older CPU.

 

[el etro]

Gamegpu.ru Mantle results

http://gamegpu.ru/action-/-fps-/-tps/battlefield-4-mantle-vs-directx-11-test-gpu.html

3970x at 5Ghz. Impossible to show Mantle gains.

Erm.... if Mantle gives better performance on the low end and on APUs that's exactly why a game geared towards the casual crowd would be using it. :hmm:

Imagine 4C apus paired with 270x...

ahahah 2 fps gain for 7950 yeah but it's true not even 20% performance gain .... dissapointed ..

A i7/i5 at 4.2Ghz is overkill to be paired with a 800-925Mhz 7950. And Mantle is still in its first phase of implement and only targeted GCN1.1 cards in this 14.1 driver.

 

[Imouto]

http://battlelog.battlefield.com/bf4/forum/threadview/2979150493821981180/1/

A lot of people on the battlelog forums stating that the Mantle renderer looks different with reduced draw distances or way more fog than DX11. Maybe the beta nature of the driver.

It's just a levels issue. No reduced draw distances conspiracy or made up stuff.

DX
Mantle
Mantle highlighted

This guy has the same issue.

DX
Mantle
Mantle highlighted

Repi acknowledged it as a bug already anyway.

But keep digging.

 

[SiliconWars]

It's just a levels issue. No reduced draw distances conspiracy or made up stuff.

DX
Mantle
Mantle highlighted

This guy has the same issue.

DX
Mantle
Mantle highlighted

Repi acknowledged it as a bug already anyway.

But keep digging.

Cool, thanks for that. This still leaves us with the issue of the Mantle setup apparently running higher settings though - look at the shadows under the roof buildings near left-side in the 2nd set of pics - that aren't shown on DX11.

I wonder if Mantle basically allows for a low level of shadowing for "free" almost and they've upped the lower settings IQ?

 

[bwat47]

Mantle is showing great potential so far!

With dx11 renderer I always drop down to 40-50 fps on certain spots on the more intensive maps like seige of shanghai. With mantle I get a MASSIVE increase in *minimum* fps.

DX11
http://s22.postimg.org/y0j8o86v3/Screenshot_Win32_0004.png [s12.postimg.org]

Mantle:
http://s22.postimg.org/qlyi9ladb/Screenshot_Win32_0005.png [s23.postimg.org]

That's about double the fps!

And this is on a 280x which apparently isn't optimized yet. I do get some annoying random stuttering with the mantle renderer though, and with the mantle render triple buffering seems to be broken (with vsync enabled it will often drop from 120 to exactly 60 fps, which is half my refresh rate, its fine with vsync disabled). I didn't get any crashes or stability issues, which suprised me. For an early beta of a new low level api it seems pretty good.