This guy reports seeing lower IQ with Mantle on
http://battlelog.battlefield.com/bf4/forum/threadview/2979150493821981180/
might be a bug or a troll, but I think it is worth looking at just to be sure Mantle isn't detrimental to the IQ
AMD Mantle API [updating]
- Thread starter MarkoIt
- Start date
This guy reports seeing lower IQ with Mantle on
http://battlelog.battlefield.com/bf4/forum/threadview/2979150493821981180/
might be a bug or a troll, but I think it is worth looking at just to be sure Mantle isn't detrimental to the IQ
Troll :/
Just look at left side building walls and rooftops, Mantle is rendering using higher settings, therefore adding additional details like shadows and better fog. Besides measuring MP map is hard as the weather changes quite quickly!
My shots show no difference but I will try to make additional comparisons later on in other spots
Quick IQ comparision:
Mantle
https://www.dropbox.com/s/we5mro05w5kudmv/Screenshot%202014-02-02%2014.51.58.png
https://www.dropbox.com/s/xvir2wlwpkh9828/Screenshot 2014-02-02 15.08.00.png
DX11
https://www.dropbox.com/s/4yvfthri2mp6ysq/Screenshot 2014-02-02 14.56.24.png
https://www.dropbox.com/s/lp2k4wrfpp9sm61/Screenshot 2014-02-02 15.04.19.png
Edit:
I take what I said above back
There are visible changes to rendering. Mainly it looks like the whole scene is brighter in Mantle. I think they just reused most of console code in Mantle which makes sense, but more comparisons is needed and I don't have time at this moment.
PS. How to do proper ScreenShot in full screen from BF4? Especially Mantle version? I'm only able to do windowed which is not ideal.
Mantle
https://www.dropbox.com/s/we5mro05w5kudmv/Screenshot%202014-02-02%2014.51.58.png
https://www.dropbox.com/s/xvir2wlwpkh9828/Screenshot 2014-02-02 15.08.00.png
DX11
https://www.dropbox.com/s/4yvfthri2mp6ysq/Screenshot 2014-02-02 14.56.24.png
https://www.dropbox.com/s/lp2k4wrfpp9sm61/Screenshot 2014-02-02 15.04.19.png
Edit:
I take what I said above back
There are visible changes to rendering. Mainly it looks like the whole scene is brighter in Mantle. I think they just reused most of console code in Mantle which makes sense, but more comparisons is needed and I don't have time at this moment.
PS. How to do proper ScreenShot in full screen from BF4? Especially Mantle version? I'm only able to do windowed which is not ideal.
Last edited by a moderator:
On my I7 2600K@ 4Ghz + HD7870@ 1100/1250 i see performance boost of 15-30% with mantle compared to DX11 (win 7). The problem is i get severe stuttering with mantle enabled. I am not even sure if word stuttering is adequate as sometime my screen freezes for 0,5s (like once every 100-200sec.). Using DX with 14.1 is much smoother but still not as good as it was with 13.12.
EDIT:
Oh and IQ seems to be the same on Mantle and DX.
EDIT:
Oh and IQ seems to be the same on Mantle and DX.
Troll :/
Just look at left side building walls and rooftops, Mantle is rendering using higher settings, therefore adding additional details like shadows and better fog. Besides measuring MP map is hard as the weather changes quite quickly!
My shots show no difference but I will try to make additional comparisons later on in other spots
Not to mention that if you take the Mantle-shot to Photoshop, you can actually get the "missing buildings" visible in it, proving they're there
overclocked_enthusiasm
Regular
Big difference in lighting/shadows in hallway. Does the game appear "brighter" in Mantle? The floor looks so much lighter. Which looks better in-game?
*Look at the tree in the center of the Vespa screenshot...much more well defined in the Mantle slide
*Look at the tree in the center of the Vespa screenshot...much more well defined in the Mantle slide
Some systems had serious stuttering issues with windows 7 in the beginning, and PCGHs nvidia system could easily be one of them. It's hard to say if they benefit from windows 8.1 or suffer from some windows 7 issue. (which has mostly been fixed since then)
Mantle gave my aging 3.8Ghz Bloomfield a much needed boost...10-15fps more on the minimal..makes me excited and angry so much performance have been left on the table for so long. The in-game stability has gotten a tad worse though....couple of game crashes so far...DX11.2 was very stable...btw did any noticed the FOV looks weird now? When turning, a vehicle at the edge of your screen-"eyes" becomes elongated! Did not noticed such strange-ness in DX11.2.
Soo...how easy is it for devs to make a Mantle version of their games? Think about it...there are now only Nvidia and AMD left, why so hard to commit to a bare-to-the-metal API? Oh, dear Dave you need to target Coh2, FF14, MGSV and WatchDogs and get 'em on board!
Soo...how easy is it for devs to make a Mantle version of their games? Think about it...there are now only Nvidia and AMD left, why so hard to commit to a bare-to-the-metal API? Oh, dear Dave you need to target Coh2, FF14, MGSV and WatchDogs and get 'em on board!
Great, thanks for the notification.
Actually it feels better in Mantle in general.
I do not get severe stuttering after disabling AV software
I got 80+ FPS with disabled Motion Blur ( \Steam\steamapps\common\Star Swarm Benchmark\Assets , open Settings_Extreme, change EnableMotionBlur=1 to 0 )
===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\hereshouldbeusername\Documents\Star Swarm\Output_14_02_02_1803.txt
Version 1.00
02/02/2014 18:03
===========================================================
== Hardware Configuration =================================
GPU: NVIDIA GeForce GTX 780
CPU: GenuineIntel
Intel(R) Core(TM) i7 CPU 940 @ 2.93GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 12874534912
Allocatable Memory: 8796092891136
===========================================================
== Configuration ==========================================
API: DirectX
Scenario: ScenarioFollow.csv
User Input: Disabled
Resolution: 1920x1080
Fullscreen: True
GameCore Update: 16.6 ms
Bloom Quality: High
PointLight Quality: High
ToneCurve Quality: High
Glare Overdraw: 16
Shading Samples: 64
Shade Quality: Mid
Deferred Contexts: Disabled
Temporal AA Duration: 16
Temporal AA Time Slice: 2
Detailed Frame Info: Off
===========================================================
== Results ================================================
Test Duration: 360 Seconds
Total Frames: 28949
Average FPS: 80.41
Average Unit Count: 4790
Maximum Unit Count: 6396
Average Batches/MS: 255.68
Maximum Batches/MS: 1078.34
Average Batch Count: 3692
Maximum Batch Count: 25452
===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\hereshouldbeusername\Documents\Star Swarm\Output_14_02_02_1803.txt
Version 1.00
02/02/2014 18:03
===========================================================
== Hardware Configuration =================================
GPU: NVIDIA GeForce GTX 780
CPU: GenuineIntel
Intel(R) Core(TM) i7 CPU 940 @ 2.93GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 12874534912
Allocatable Memory: 8796092891136
===========================================================
== Configuration ==========================================
API: DirectX
Scenario: ScenarioFollow.csv
User Input: Disabled
Resolution: 1920x1080
Fullscreen: True
GameCore Update: 16.6 ms
Bloom Quality: High
PointLight Quality: High
ToneCurve Quality: High
Glare Overdraw: 16
Shading Samples: 64
Shade Quality: Mid
Deferred Contexts: Disabled
Temporal AA Duration: 16
Temporal AA Time Slice: 2
Detailed Frame Info: Off
===========================================================
== Results ================================================
Test Duration: 360 Seconds
Total Frames: 28949
Average FPS: 80.41
Average Unit Count: 4790
Maximum Unit Count: 6396
Average Batches/MS: 255.68
Maximum Batches/MS: 1078.34
Average Batch Count: 3692
Maximum Batch Count: 25452
===========================================================
Last edited by a moderator:
DavidGraham
Veteran
Turns out it is a bug, Repi says they will fix it.Quick IQ comparision:
I take what I said above back
There are visible changes to rendering. Mainly it looks like the whole scene is brighter in Mantle. I think they just reused most of console code in Mantle which makes sense, but more comparisons is needed and I don't have time at this moment.
https://twitter.com/repi/status/429985248995467265
In case anyone encounters the same thing, Mantle requires Intel integrated GPU to be disabled in device manager to work. Star Swarm would crash every time on Mantle version otherwise.
i5-3570k @ 4.2 + 7870 @ 1125/1300
1080p 28 average fps vs 38 average fps
It's a massive difference visually during the stress test. The D3D version has several slide shows which Mantle seems to handle with no problems.
Interestingly at 5040x1040 Eyefinity, if I turn off motion blur I get 1fps average difference between D3D and Mantle. Probably because I'd be so GPU bound to make a difference.
i5-3570k @ 4.2 + 7870 @ 1125/1300
1080p 28 average fps vs 38 average fps
It's a massive difference visually during the stress test. The D3D version has several slide shows which Mantle seems to handle with no problems.
Interestingly at 5040x1040 Eyefinity, if I turn off motion blur I get 1fps average difference between D3D and Mantle. Probably because I'd be so GPU bound to make a difference.
Last edited by a moderator:
My above post was actually on ...BF4..
Now that it is mentioned Star Swarm benchmark do not need a registration, i just tried it and got about 90fps without motion blur with Mantle....your GTX 780 is 80fps...that is impressive, better than my DX11 score which was only 65fps!
Good to hear
UniversalTruth
Veteran
Can anyone test the cards with Core 2 Quad? It should give those processors new life, no?
Not sure that's the case. From all reports, porting over to mantle isn't a huge investment in resources. These days engine developers are designing games to work on windows and Android OS which is on a different API and a different OS. Mantle seems like a pretty easy switch in comparison..but I'm not a developer...so could be way off base.
Yeah...I imagine that soon there will be a few killer apps for mantle that show off how much better games can be with so many limits removed!
DavidGraham
Veteran
Speaking of which, so far Mantle's main advantage has been CPU centric, I still don't get why it's limited to only GCN? why can't it be extended to prior architectures?
I mean with PhysX and CUDA, it's pretty straightforward. it's GPU centric and and it can run on anything from G80 (the first unified design to allow for it) up to K110 .. Mantle on the other hand is both GPU and CPU, at least the CPU portion of it should be available for other architectures.
I mean with PhysX and CUDA, it's pretty straightforward. it's GPU centric and and it can run on anything from G80 (the first unified design to allow for it) up to K110 .. Mantle on the other hand is both GPU and CPU, at least the CPU portion of it should be available for other architectures.
