The AMD Mantle Thread

[AtenRa]

Core i7 3770K 4GHz
HD7950 1GHz
Win 8.1 64bit
Cat 4.1

Dragon Pass 64 player

1080p custom settings.

D3D 11.1

Mantle

 

[krumme]

Core i7 3770K 4GHz
HD7950 1GHz
Win 8.1 64bit
Cat 4.1

Dragon Pass 64 player

1080p custom settings.

D3D 11.1

Mantle

Its because you play sniper. Mantle have a build in punishment for you. With 14.2 beta you will get -25% xp constantly.

 

[krumme]

Its because you play sniper. Mantle have a build in punishment for you. With 14.2 beta you will get -25% xp constantly.

Well more seriously i havnt seen anything like that. I play ultra 1080 without msaa and post aa. And my experience i far more smooth but i have a bug with hickups every 5 min or so. 3570 at 4ghz 7970 1 ghz 16 ram win7.

 

[AtenRa]

Its because you play sniper.

I barely play sniper, just picked that kit for the pics.

 

[AtenRa]

Well more seriously i havnt seen anything like that. I play ultra 1080 without msaa and post msaa. And my experience i far more smooth but i have a bug with hickups every 5 min or so. 3570 at 4ghz 7970 1 ghz 16 ram.

Than the pics above ???

 

[krumme]

I barely play sniper, just picked that kit for the pics.

Do you get the same difference when setting terrain high/ultra?

 

[AtenRa]

Do you get the same difference when setting terrain high/ultra?

Havent used that, ill try tomorrow

 

[krumme]

Havent used that, ill try tomorrow

Lol sorry. Misunderstanded you. But would like to know results there also.

 

[AtenRa]

Lol sorry. Misunderstanded you. But would line to know results there also.

I mean i havent changed the Terrain settings yet, ill try terrain on ultra tomorrow.

 

[krumme]

Btw interesting to know difference between 3570 and 3770 and the effect of ht with mantle. Anyone knows?

 

[AtenRa]

Btw interesting to know difference between 3570 and 3770 and the effect of ht with mantle. Anyone knows?

Yea i was planing of disabling HT and see how it will affect the game with and without mantle.

 

[krumme]

Yea i was planing of disabling HT and see how it will affect the game with and without mantle.

Super.

 

[krumme]

Btw do you experience hickups?

 

[OCGuy]

Back to 13.12 drivers for now. Much too buggy, but FPS gains were seen. I just dont know that 10 frames is worth the trouble when triple digits are flashing in the corner.

The point of Mantle is obviously not for FPS games and people with decent hardware. If there was a AAA RTS title that could have served as the proof of concept, Im sure AMD would have jumped on it.

Creating a massive battle with hundreds of units is a much more established way of creating a bottleneck than outlier hardware scenarios and 64 player environments that can't be replicated.

In a RTS, we could see what Mantle actually adds to the experience. The ability to have a higher unit cap is tangible. Going from 150 to 170 FPS in a shooter is only useful in forum wars.

Even in these hand-picked perfect situations, nothing I've seen goes beyond what any strong driver release could accomplish. Those gains are not going to hold up against other players' long-known plans to do something very similar to Mantle.

Microsoft announced DX11.2 "close to metal" API tweaks in June, and a Nv has had plans for the extra processor on Maxwell for years.

AMD actually has one more release in the near future that is meant to compliment Mantle. TrueAudio isnt about sound being "better", it is about removing audio processing burden off the CPU.

Fun times ahead, no need for hyperbole or poor testing methods.

 

[Stuka87]

Has anybody seen any benchmarks that don't use $1000 CPU's? Majority of gamers use either mid range CPU's (FX or i5). Or still have older CPU's.

Would love to see comparisons with this set up.

 

[AtenRa]

Btw do you experience hickups?

yes, when i set an fps cap at 60, when i remove the fps cap i barely have any CPU spikes.

 

[OCGuy]

So is it worth me trading up a pair of Titans and a 3930K to get Mantle?

That is a downgrade.

 

[Stuka87]

That is a downgrade.

I think you mean side grade. As the cards trade blows depending on the game.

 

[jpiniero]

The point of Mantle is obviously not for FPS games and people with decent hardware. If there was a AAA RTS title that could have served as the proof of concept, Im sure AMD would have jumped on it.

I think it was really meant for next-gen (ie: designed for PS4/XB1) titles. Those titles will obviously be more demanding, and using Mantle might make those titles playable on older PC hardware moreso than if they had just stuck with D3D.

 

[krumme]

Back to 13.12 drivers for now. Much too buggy, but FPS gains were seen. I just dont know that 10 frames is worth the trouble when triple digits are flashing in the corner.

The point of Mantle is obviously not for FPS games and people with decent hardware. If there was a AAA RTS title that could have served as the proof of concept, Im sure AMD would have jumped on it.

Creating a massive battle with hundreds of units is a much more established way of creating a bottleneck than outlier hardware scenarios and 64 player environments that can't be replicated.

In a RTS, we could see what Mantle actually adds to the experience. The ability to have a higher unit cap is tangible. Going from 150 to 170 FPS in a shooter is only useful in forum wars.

Even in these hand-picked perfect situations, nothing I've seen goes beyond what any strong driver release could accomplish. Those gains are not going to hold up against other players' long-known plans to do something very similar to Mantle.

Microsoft announced DX11.2 "close to metal" API tweaks in June, and a Nv has had plans for the extra processor on Maxwell for years.

AMD actually has one more release in the near future that is meant to compliment Mantle. TrueAudio isnt about sound being "better", it is about removing audio processing burden off the CPU.

Fun times ahead, no need for hyperbole or poor testing methods.

Bs rant that spells "anxiety". We play 64 player all the time. Who cares if it can be replicated the same way. We game. We dont bm.

No way are you constantly having 3 digits playing bf4 on dx. What matters is getting high fps all the time in a smooth gameplay. Mantle does and will do that even better. Ofcource its a beta and it shows especially for the installation issues. For a first time api and driver its generally damn bug free.

 

[OCGuy]

I think it was really meant for next-gen (ie: designed for PS4/XB1) titles. Those titles will obviously be more demanding, and using Mantle might make those titles playable on older PC hardware moreso than if they had just stuck with D3D.

This makes no sense. Consoles are never "next gen", except in the context of their own world.

 

[krumme]

Has anybody seen any benchmarks that don't use $1000 CPU's? Majority of gamers use either mid range CPU's (FX or i5). Or still have older CPU's.

Would love to see comparisons with this set up.

There is plenty to be had with a i5 4c 4t be it sb ib or hw. They are exactly perf wise where you will feel it a lot when gaming unless you are not sensitive to it or game in the 40 fps range. We have i5 4c 4t ib and hw in house with gcn and its just giving a more fluid gameplay. Reminds me of consoles. I dont give a damn about assessing fps. Stopped doing that 10 years ago. Now i just trust my experience.

 

[f1sherman]

Ofcource its a beta and it shows especially for the installation issues. For a first time api and driver its generally damn bug free.

Bug free??

Lets see:
Mantle hickups(rare but very pronounced), performance degradation(1 single game that I bothered to check - Insurgency, others are reporting similar issues with other games),
artifacting on Desktop(weird blocks on UI), PowerPlay control lost (except in CCC), RPM sensor lost.
And all this in 20 minutes of casual PC usage, nothing fancy.

Enough for me. Back to 13.12, although CCC reports it as 14.1 LMAO.

Mantle definitely has huge upside potential. But, bug free...

 

[Shivansps]

Im sorry but i dont belive the visual quality loss is a bug, its a bug because people figured it out, this is the exact same optimization strategy used in consoles that Mantle was claiming that whould made possible, its not just the fog, even it you belive that the fog is a bug.

DICE warned before hand that Mantle whould have less quality because of bugs? reviews did see to have no idea about it.

 

[SiliconWars]

Im sorry but i dont belive the visual quality loss is a bug, its a bug because people figured it out, this is the exact same optimization strategy used in consoles that Mantle was claiming that whould made possible, its not just the fog, even it you belive that the fog is a bug.

DICE warned before hand that Mantle whould have less quality because of bugs? reviews did see to have no idea about it.

Or maybe it's a bug?

http://forums.anandtech.com/showpost.php?p=36023377&postcount=123