Mantle Feedback/Bugs

LtMatt · 2 Feb 2014 at 09:59

Please use this thread to discuss Mantle, give feedback and discuss or report bugs/problems.

The old Mantle thread can be found here.

AMD Catalyst™ 14.1 Beta Driver for Windows®

Download Link
http://support.amd.com/en-us/kb-articles/Pages/latest-catalyst-windows-beta.aspx

Known issues list for AMD Catalyst™ 14.1

Last Updates: 2/1/2014
Article Number: NA
Below are the known issues AMD has found in the first Mantle-enabled Beta driver. We are providing this driver as is to highlight the potential of Mantle and are hard at work to further enhance the user experience.
Intermittent stuttering or stability issues may occur when utilizing Mantle with AMD CrossFire™ technology in BattleField 4™
Mantle performance for the AMD Radeon™ HD 7000/HD 8000 Series GPUs and AMD Radeon™ R9 280X and R9 270X GPUs will be optimized for BattleField 4™ in future AMD Catalyst™ releases. These products will see limited gains in BattleField 4™ and AMD is currently investigating optimizations for them.
Multi-GPU support under DirectX® and Mantle will be added to StarSwarm in a future application patch
Notebooks based on AMD Enduro or PowerXpress™ technologies are currently not supported by the Mantle codepath in Battlefield 4™
AMD Eyefinity configurations utilizing portrait display orientations are currently not supported by the Mantle codepath in Battlefield 4™
Graphics hardware in the AMD A10-7850K and A10-7700K may override the presence of a discrete GPU under the Mantle codepath in Battlefield 4™
AMD testing for the AMD Catalyst™ 14.1 Beta has been concentrated on the following products: AMD Radeon™ R9 290X, R9 290, R9 280, R9 270, R7 260X, R7 260, HD 7000 Series, HD 8000 Series, A10-7850K and A10-7700K. Future AMD Catalyst™ releases will include full test coverage for all AMD products supported by Mantle.
If you encounter an issue not identified on this list, please visit www.amd.com/betareport to submit a bug report. We appreciate your help in shaping the future of Mantle!

Feature Highlights of The AMD Catalyst™ 14.1 Beta Driver for Windows
Support for the following new AMD Desktop APU (Accelerated Processors) products:
AMD A10-7850K
AMD A10-7700K
Mantle Beta driver
AMD's Mantle is a groundbreaking graphics API that promises to transform the world of game development to help bring better, faster games to the PC
Performance gain of up to 45%(versus the DirectX version) for Battlefield 4 on the R9 290 Series
Performance gain of up to 200% (versus the DirectX version) for Star Swarm on the R9 290 Series
AMD Catalyst 14.1 Beta must be used in conjunction with versions of these applications that support Mantle
It is expected that these applications will have future updates to support additional AMD Mantle features
AMD Mantle Beta driver is currently supported on:
AMD Radeon™ R9 Series GPUs
AMD Radeon™ R7 Series GPUs
AMD Radeon™ HD 7000 Series GPUs
AMD Radeon™ HD 8000 Series GPUs
AMD A10-7000 Series and AMD A8-7000 Series APUs
For additional details please see the AMD Mantle Technology FAQ on amd.com
Enhanced AMD CrossFire frame pacing – Support for 4K panel and Eyefinity non-XDMA CrossFire solutions (including the AMD Radeon R9 280, 270 Series, 7900 Series, 7800 Series) and Dual Graphics configurations
Frame pacing ensures that frames rendered across multiple GPUs in an AMD CrossFire configuration will be displayed at an even and regular pace
Supported on 4K panels and Eyefinity configurations
Supported on AMD Dual Graphics configurations
Supported on DirectX® 10 and DirectX 11 applications
Resolved issue highlights of AMD Catalyst 14.1 Beta
Resolves ground texture flickering seen in Total War: Rome 2 with high settings (and below) set in game
Resolves flickering texture corruption when playing Call of Duty: Ghosts (multi-player) in the space station level

Resolved Issues

Ground texture flickering seen in Total War: Rome 2 with high settings (and below) set in game
Flickering texture corruption when playing Call of Duty: Ghosts (multi-player) in the space station level
Blu-ray playback using PowerDVD black screen on extended mode
Streaming VUDU HD/HDX content on Sharp PN-K321 (DP) causes the right-side half to flicker in and out
Black screen happened after wake up the monitor
Full screen issue at rotation in DX9 mode
Video window black screen when using Samsung Kies to play video
Crysis2 negative scaling in outdoor scene
Crysis2 has insufficient CrossFire scaling in some scene
Red Faction: The game has no or negative crossfire scaling with DX9 and DX11
Age of Conan has corruption and performance issues with crossfire enabled
Company of Heroes shadows are corrupted when using crossfire
Resident Evil5 's performance is unstable when display mode set to Window mode
Total War: Shogun 2 flickering menu/text
Frame rate drop when disabling post-processing in 3DMark06
Negative Crossfire scaling with game "The Secret World" in DX11 mode
F1 2012 Crashes to desktop
Tomb Raider Hair Simulation Stutters on CFX
Negative CrossFire scaling experienced in Call of Duty
Battlefield 3 performance drop on Haswell systems
Choppy video playback on 4k Video
VSync ON Tearing with 2x1 Eyefinity SLS CrossFire
Far Cry 3 - Game flickering while changing resolutions
Display corruption and BSOD occurs when extending a display after disabling Multiple GPU SLS array
Flickering seen when enable three 4kx2k panels at the same time
No Video, just a black screen when setting Chrome to run in "High Performance" when playing certain video clips
Image crashed on Starcraft game

LtMatt · 2 Feb 2014 at 10:00

Useful In Game Commands To Monitor Performance

Bring up the console in game.

FPS Monitor type - PerfOverlay.DrawFps 1

Cpu+Gpu performance overlay type - PerfOverlay.DrawGraph 1

To take a screenshot type - Render.Screenshot

To enable a FPS log type at start of bench - PerfOverlay.FrameFileLogEnable 1

To stop FPS log at end of bench type - PerfOverlay.FrameFileLogEnable 0

A little tip. Whilst the level/map is loading you can type in PerfOverlay.FrameFileLogEnable 0 and press enter. Then when you need to start logging just bring up the console, press down arrow key on your keyboard or S and that last written command comes up. Saves you retyping it each time. Just change the 0 to 1 and hit enter to start recording.

GefGz · 2 Feb 2014 at 10:04

Oh. They released it when I went to work

- I'll edit when I can install that driver!

ICDP · 2 Feb 2014 at 10:08

DX - Run 1: Scenario RTS
FPS:10.76

Oxide Games
Star Swarm Stress Test - ©2013
Star Swarm\Output_14_02_01_1645.txt
Version 1.00
02/01/2014 16:45
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 17121140736
Allocatable Memory: 140737488224256
===========================================================

== Configuration ==========================================
API: DirectX
Scenario: ScenarioRTS.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: 3875

Average FPS: 10.76
Average Unit Count: 4019
Maximum Unit Count: 5577
Average Batches/MS: 1088.16
Maximum Batches/MS: 1576.31
Average Batch Count: 98692
Maximum Batch Count: 165968
===========================================================

With DX - Run 2: Scenario RTS
FPS:11.14

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
Star Swarm\Output_14_02_02_0920.txt
Version 1.00
02/02/2014 09:20
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 17121140736
Allocatable Memory: 140737488224256
===========================================================

== Configuration ==========================================
API: DirectX
Scenario: ScenarioRTS.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: 4012

Average FPS: 11.14
Average Unit Count: 3811
Maximum Unit Count: 5504
Average Batches/MS: 1069.69
Maximum Batches/MS: 1582.65
Average Batch Count: 94210
Maximum Batch Count: 174755
===========================================================

Mantle - Run 1: Scenario RTS
FPS:29.27

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
Star Swarm\Output_14_02_02_0926.txt
Version 1.00
02/02/2014 09:26
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 17121140736
Allocatable Memory: 140737488224256
===========================================================

== Configuration ==========================================
API: Mantle
Scenario: ScenarioRTS.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: 10537

Average FPS: 29.27
Average Unit Count: 4174
Maximum Unit Count: 5471
Average Batches/MS: 2501.25
Maximum Batches/MS: 3425.00
Average Batch Count: 85354
Maximum Batch Count: 175015
===========================================================

167.31% increase performance in the exact same scenario. Excellent stuff from just one test. I don't have DF4 (and have no intention of getting it). I can't wait for Thief 4.

Intel i7 4770k @ 4.4 GHz
R9 290X at 1100/1300
Windows 8.1
16GB DDR3 2200

Rofflay · 2 Feb 2014 at 10:08

Matt i think with regards to this place being the place it is i do not think too many people will be interested to hear about anything other than how it does in the biggest games like Battlefield 4 and with equipment most of us use.

Example would be i5 v i7 with regards to Hyperthreading and Mantle and how it effects perhaps clock speeds and overclocks? Is it really that hot for people who lack a little bit of cpu grunt in BF4 64p MP? And just how much percentage increase can people expect on a 290X?

LtMatt · 2 Feb 2014 at 10:12

Yes this is for posting of benchmarks and stuff only really. I will let people post them and then try and find some way of adding them coherently to the first two posts. Sorry for the late posting of the thread. I'm amazed no one else did it. I was away last night getting drunk round a friends, came home and went straight to best. Woke up this morning with the worst hangover ever, but seeing this has cured it somewhat. Now to install the drivers.

Sacrifice · 2 Feb 2014 at 10:22

LtMatt said:
Yes this is for posting of benchmarks and stuff only really. I will let people post them and then try and find some way of adding them coherently to the first two posts. Sorry for the late posting of the thread. I'm amazed no one else did it. I was away last night getting drunk round a friends, came home and went straight to best. Woke up this morning with the worst hangover ever, but seeing this has cured it somewhat. Now to install the drivers.

I must warn you, my crossfire setup was seeing constant up and down from 30 to 20, the perfoverlay graph looked like one solid line of green! Also, if you haven't already, disable the igpu in device manager, otherwise clicking options in bf4 crashes for some reason.

Beerbaron · 2 Feb 2014 at 10:36

What are you using to test/capture those stats?

LtMatt · 2 Feb 2014 at 10:40

Sacrifice said:
I must warn you, my crossfire setup was seeing constant up and down from 30 to 20, the perfoverlay graph looked like one solid line of green! Also, if you haven't already, disable the igpu in device manager, otherwise clicking options in bf4 crashes for some reason.

Installed with no issue so far, but ill keep my eyes peeled. Using the msi afterburner overlay causes the game to crash, so keep the overlay disabled/msi afterburner shut down for now.

Beerbaron said:
What are you using to test/capture those stats?

Place this file in your Battlefield 4 main directory with the exe. This will put up render graphs and an fps counter in game. It will also output render times and fps details to My documents Battlefield 4 folder.

http://www.sendspace.com/file/wbc2bd

Calneon · 2 Feb 2014 at 10:45

ICDP said:
DX - Run 1: Scenario RTS
FPS:10.76

Oxide Games
Star Swarm Stress Test - ©2013
Star Swarm\Output_14_02_01_1645.txt
Version 1.00
02/01/2014 16:45
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 17121140736
Allocatable Memory: 140737488224256
===========================================================

== Configuration ==========================================
API: DirectX
Scenario: ScenarioRTS.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: 3875

Average FPS: 10.76
Average Unit Count: 4019
Maximum Unit Count: 5577
Average Batches/MS: 1088.16
Maximum Batches/MS: 1576.31
Average Batch Count: 98692
Maximum Batch Count: 165968
===========================================================

With DX - Run 2: Scenario RTS
FPS:11.14

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
Star Swarm\Output_14_02_02_0920.txt
Version 1.00
02/02/2014 09:20
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 17121140736
Allocatable Memory: 140737488224256
===========================================================

== Configuration ==========================================
API: DirectX
Scenario: ScenarioRTS.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: 4012

Average FPS: 11.14
Average Unit Count: 3811
Maximum Unit Count: 5504
Average Batches/MS: 1069.69
Maximum Batches/MS: 1582.65
Average Batch Count: 94210
Maximum Batch Count: 174755
===========================================================

Mantle - Run 1: Scenario RTS
FPS:29.27

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
Star Swarm\Output_14_02_02_0926.txt
Version 1.00
02/02/2014 09:26
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 17121140736
Allocatable Memory: 140737488224256
===========================================================

== Configuration ==========================================
API: Mantle
Scenario: ScenarioRTS.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: 10537

Average FPS: 29.27
Average Unit Count: 4174
Maximum Unit Count: 5471
Average Batches/MS: 2501.25
Maximum Batches/MS: 3425.00
Average Batch Count: 85354
Maximum Batch Count: 175015
===========================================================

167.31% increase performance in the exact same scenario. Excellent stuff from just one test. I don't have DF4 (and have no intention of getting it). I can't wait for Thief 4.

Intel i7 4770k @ 4.4 GHz
R9 290X at 1100/1300
Windows 8.1
16GB DDR3 2200

That benchmark really is meaningless in terms of actual gaming performance. It's tailor made to make Mantle look amazing.

=XDC=FluphyBunny · 2 Feb 2014 at 10:51

Guys and gals when you post your results don't forget to add your gpu/cpu specs plus overclocks

EL_JOCK · 2 Feb 2014 at 10:55

do we have a planned bench with BF4 i.e single player with x level or Mp with x map?

ICDP · 2 Feb 2014 at 10:56

Calneon said:
That benchmark really is meaningless in terms of actual gaming performance. It's tailor made to make Mantle look amazing.

Scenario A: Oxide decided to deliberately cripple DX on a game engine they are using for their own games? Knowing full well it would render Nvidia cards worthless thus alienating the majority of their potential market.

Scenario B: Oxide used Mantle to enable a level of detail and graphical quality that is just not possible on DX?

I'm going with scenario B thanks.

LtMatt · 2 Feb 2014 at 11:03

EL_JOCK said:
do we have a planned bench with BF4 i.e single player with x level or Mp with x map?

I will upload one for everyone to copy. I'm working on it now. Its a very consistent and easy to bench scene.

Calneon · 2 Feb 2014 at 11:03

ICDP said:
This is bullcrap. Sorry, I'm, sick of reading that tripe. So Oxide decided to deliberately cripple DX on a game engine they are using for their own games? Knowing full well it would render Nvidia cards worthless thus alienating the majority of their potential market.

Does that really make sense to you?

I'm not saying their engine cripples DX or anything like that. It would be the same for every engine. The scene is designed to maximise the performance boost from Mantle, i.e. having thousands of ships (draw calls) on the screen at once. No actual game would have that many draw calls in a scene at the same time. Mantle increases performance by removing CPU bottlenecks, so they create a benchmark which maximises the CPU bottleneck... Make sense?

Why do you think people are only seeing ~10% improvements in BF4 and 50%+ improvements in this benchmark?

BuzzLloydyear · 2 Feb 2014 at 11:12

Done 3 runs of Star Swarm on "Follow" mode and "Extreme" settings for each mode. I then averaged the 3 results. To be honest I could probably do with more runs than 3, but it takes a long time. I like that the differing performances gives you a more live gaming scenario and not just the same point with the same bottlenecks at the same point. I think it gives a better overall idea of performance, when you average over multiple runs.

The following results were on my 7990 (stock clocks) with the 4670k (4.6ghz).

== Mantle Average Results (3 Runs) =======================[
Test Duration: 360 Seconds
Total Frames: 18,848.66

Average FPS: 52.35
Average Unit Count: 4,487
Maximum Unit Count: 5,778
Average Batches/MS: 826.39
Maximum Batches/MS: 3,225
Average Batch Count: 17,820
Maximum Batch Count: 114,226
==================================================

== Direct3D Average Results (3 Runs) =====================
Test Duration: 360 Seconds
Total Frames: 14,579.66

Average FPS: 40.50
Average Unit Count: 4,106
Maximum Unit Count: 5,555
Average Batches/MS: 617.45
Maximum Batches/MS: 1,275
Average Batch Count: 17,244
Maximum Batch Count: 113,443
==================================================

What I was surprised to see was Direct3D max fps absoloutely smashed Mantle's max fps. I am talking by close to 100 fps (In quiet scenes, Mantle maxes out at around 75 fps, Direct3D went to around 170). However the lows on Direct3D were insanely low (around 7 fps) compared to Mantles (around 25fps). Overall Mantle was faster, pushing out more batches and generally better. I don't quite understand the differences in the max fps, but I am sure after many more runs I will start to get an idea with it.

Stanners · 2 Feb 2014 at 11:13

ICDP said:
Scenario A: Oxide decided to deliberately cripple DX on a game engine they are using for their own games? Knowing full well it would render Nvidia cards worthless thus alienating the majority of their potential market.

Scenario B: Oxide used Mantle to enable a level of detail and graphical quality that is just not possible on DX?

I'm going with scenario B thanks.

Arent peple only seeing 50% gpu usage under dx, somethings not right is it.

Also your ninja edit was caught which suggests you are to close and emmotional to hold an unbiased view on the subject.

ICDP · 2 Feb 2014 at 11:13

Calneon said:
I'm not saying their engine cripples DX or anything like that. It would be the same for every engine. The scene is designed to maximise the performance boost from Mantle, i.e. having thousands of ships (draw calls) on the screen at once. No actual game would have that many draw calls in a scene at the same time. Mantle increases performance by removing CPU bottlenecks, so they create a benchmark which maximises the CPU bottleneck... Make sense?

Why do you think people are only seeing ~10% improvements in BF4 and 50%+ improvements in this benchmark?

Sorry, I edited my post as I felt it was far too confrontational in the format you quoted. My apologies for that.

To answer your question BF4 and Star Swarm uses massively different graphics engines. The reason for this massive difference between DX and Mantle in Star Swarm is explained by Oxide in this video.

http://www.youtube.com/watch?v=6PKxP30WxYM

chief barker · 2 Feb 2014 at 11:14

I am to put simply blown away by Mantle.

I've gained about 50 FPS

Exactly the same settings as last night/same clock speeds.

The only issue i've got is all of a sudden my FPS drops to 30 (hitching) but I'm sure it will get fixed with a future driver update.

Edit.

My 7970s are working harder too, the fans are going crazy

Edit again,

IT'S BEAUTIFUL

=XDC=FluphyBunny · 2 Feb 2014 at 11:18

edit: meant to quote ltmatt

I will upload one for everyone to copy. I'm working on it now. Its a very consistent and easy to bench scene.

To be honest the only useful benchmark for bf is multiple timed plays in multiplayer. Doesn't matter what its like in single player unless that benchmark mirrors a multiplayer style scenario? (unlikely).

Best bet play on a full 48/64 player server in DX. Play again in Mantle. Multiple timed benchmarks and compare results. Its a faff on but benching always is.