AMD Catalyst 14.6 brings mantle to notebooks

Didn't see this posted so I thought I'd share:

AMD Catalyst 14.6 brings Eyefinity with mixed resolutions support | VideoCardz.com

Supposedly available tomorrow!

Sent from Tapatalk from an increasingly expensive smartphone

EDIT: Improvements include Eyefinity enhancements, 25%-28% better performance in Watch Dogs as well as better performance in other games like Thief, mantle to notebooks.

 

It's about time - notebook gamers should not be treated marginally by AMD and Mantle has been out on the market ~6 months already.
It would be nice to see a decent performance increase in Thief.

 

Well, if this article is not a big fat lie, that wish just may be granted. We shall see my friend!

 

nice! would love to give it a try but i don't have an amd gpu

 

Well, with an overclocked GTX780M in your lappy I think you have nothing to regret... ;-)

 

i guess bf4 would run better in mantle as well.

 

I'm interested to see if AMD is going to list the R9 M290X as officially supported by the 14.6 driver.

 

Guru3D offers the download link :thumbsup:

 

Here's the download link from Guru3d

Thanks for the heads up Tornator! Hopefully it'll be up on AMD's site very soon.

EDIT: If you download this, make sure you download the SECOND option for the mobile drivers...

 

Downloaded 14.6 from Guru3d and running Watch Dogs all on High settings with FXAA, runs very smooth and nice!
Great game so far.

 

14.6 beta driver is up on amd's website.

http://www2.ati.com/drivers/beta/amd-catalyst-14.6-beta-v1.0-windows-may23.exe

 

I think that notebook users see very little gains with Mantle except in some few CPU bound games. The CPUs are just too powerful for the GPUs we use.
But I would love to see a review with DX vs Mantle

 

Yeah, I think you're going to be proven correct Mr. Fire of the Clouds. Everything I've read/seen has shown it to be of great benefit to lesser CPU's and not much at all to the powerful i7's.

Sent from Tapatalk from an increasingly expensive smartphone

 

Say hi to MSI GX users

 

My benching on Fire Strike increased by 2% and on 3DMark11 I got an increase of almost 5% with 14.6. :thumbsup:

 

I tried using mantle on BF4.. I'm getting lagg.. Any idea why? DX11 works perfectly

 

I don't have any games that work with Mantle so I can't test it yet. I went ahead and grabbed the drivers straight from AMD and I ended up getting another small little boost in my benchmarks. If you downloaded the one from guru3d I'd try the official one. Although I'm not positive there's any difference at all with the two...

AMD Catalyst 14.6 Beta drivers

 

I'm getting severe lag on BF4 with Mantle as well, when DX11 doesn't lag (at the exact same settings); all done with the AMD Catalyst 14.6 drivers

I did notice that the VRAM usage (in MSI Afterburner) is ~ 1.7 GB @ 1920x1080, High settings, 100% resolution scale, in Mantle mode, well over the 1 GB VRAM of my card. If I drop the resolution scale to < 40%, the game stutters way less, VRAM usage goes way down, just a hair below 1 GB.

So, clearly something is wrong with the drivers; severe memory leakage or something like that when Enduro and Mantle are both used.

My Desktop PC with the HD 7950 runs BF4 @ Ultra in Mantle mode perfectly fine, zero stutter.

 

can someone show their benchies before and after?

 

That probably explains it...

Additionally, it seems AMD's website is broken.. For any drivers, the page just loads and there's random crap there....

 

Fire Strike before:

View attachment 112625

Fire Strike after:

View attachment 112626

3DMark11 before:

View attachment 112627

3DMark11 after:

View attachment 112628

 

Does 3DMark support Mantle? I don`t think so
You did have some extra increase in the benchmark though. Maybe its just the driver itself doing some magic?

You got a minor 150 points increase in graphics in both benchmarks.
CPU score (Physics score) went up by a good 13% in 3DMark11.

But even the slightest change in CPU speed will reflect on better benchmark score even though you won`t notice anything in games.
Did you notice better FPS in the games that support Mantle?

 

I just have the basic version of 3DMark so I can't choose anything other than to run the full tests, so I'm not sure. I also don't have any games that support Mantle so I can't speak of that either. Wow...typing that I sound pretty useless

 

For the purpose of testing and benchmarking Mantle vs DX you can download StarSwarm Stress Test in Steam. It's free...

 

You sir are a gentleman and a scholar! I'll be checking out StarSwarm soon!

Sent from Tapatalk from an increasingly expensive smartphone

 

OK...let me preface this by saying HOLY CRAP, I DON'T EVEN BELIEVE THESE BENCHMARKS!!! I took Marecki's advice and download StarSwarm Stress Test that has both DX and Mantle API to bench with. I'm just going to post these benchies:

DX:
===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\lsaintignan\Documents\Star Swarm\Output_14_05_29_2247.txt
Version 1.10
05/29/2014 22:47
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon(TM) HD8970M
CPU: GenuineIntel
Intel(R) Core(TM) i7-4700MQ CPU @ 2.40GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 17099153408
Allocatable Memory: 140737488224256
===========================================================


== 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: 9113

Average FPS: 25.31
Average Unit Count: 4068
Maximum Unit Count: 5460
Average Batches/MS: 463.51
Maximum Batches/MS: 1207.36
Average Batch Count: 21533
Maximum Batch Count: 122205
===========================================================

Mantle:

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\lsaintignan\Documents\Star Swarm\Output_14_05_29_2254.txt
Version 1.10
05/29/2014 22:54
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon(TM) HD8970M
CPU: GenuineIntel
Intel(R) Core(TM) i7-4700MQ CPU @ 2.40GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 17099153408
Allocatable Memory: 140737488224256
===========================================================


== Configuration ==========================================
API: Mantle
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: 13308

Average FPS: 36.97
Average Unit Count: 4262
Maximum Unit Count: 5389
Average Batches/MS: 814.99
Maximum Batches/MS: 7774.43
Average Batch Count: 24114
Maximum Batch Count: 114489
===========================================================

Ok...so using Mantle API, according to this StarSwarm stress test, my average FPS went from 25.31 to 36.97...that's a friggin' 46.1% jump!! This can't be right. I even thought Enduro was throwing it off and activating Intel Graphics instead of my AMD card on the DX, so I turned off my Intel Graphics. I also ran the test multiple times with basically the same results.

If this is true, and Mantle gives this much of a boost...like I said at the beginning...HOLY CRAP!!

What do you all make of these numbers?

Also here's a grab of my sensors during the Mantle test:

 

That's in line with what others have reported. In fact, I've seen many people report even greater gains than you, up to 100% increase in Mantle over DX11. Due to its nature, Star Swarm really is a worst-case scenario for DX, which is why Mantle provides such a significant performance improvement.

 

Yeah, DX was struggling during the most intense moments really bad. Mantle just trucked along. I'm glad to hear others are seeing good gains as well.

 

That's because StarSwarm was designed to depict strengths of Mantle. It's not going to be such a drastic improvement in every game which uses Mantle though.

 

So basically this is the best case scenario.

 

Don't know if you noticed this, but load is near identical on every single thread. You won't ever find this kind of perfect multi-core CPU utilization in any DX game, only Mantle. :thumbsup:

Yes.

 

I did notice this and it made me smile

 

Does 7970m support mantle?


Sent from my iPod touch using Tapatalk

 

Yup it does.. Luckily unlike the 6000 series... But it seems Mantle is broken for me in BF4.. So much lag in comparison to DX11

 

I am sure mantle will improve with consequent driver updates. Good news really. Seems like my 920xm won't be a bottleneck for quite some time if mantle is adopted by all games.

 

I never tried watch dogs with 14.4 but I'm very happy with the performance on 14.6 on high, high textures, fxaa, motion blur turned off.

I don't think 920/940xm bottlenecks even at 3.1ghz. Load is about 70-80% across all threads with GPU usage in 90s.

Sent from my One S using Tapatalk

 

BF4 just had an update yesterday and mantle seems to work much better... less lagg but DX11 is still faster for me...

 

That is pretty much bottlenecked. At times does the load fluctuate to 100%? Astounding that even with mantle the old chip is really starting to show it's age! You had better get that voltmod done

 

Nah no fluctuation. I can increase load by turning off v-sync but I prefer to keep it on.

Sent from my One S using Tapatalk

 

Hi guys,

I have a samsung series 3 laptop with these specs:
Intel Core i5-3210M (2.5 GHz), Amd Radeon HD8750M (1GB gDDR3 VRAM), 8 GB RAM

I downloaded 14.6 driver because i wanted to try out mantle on Star Swarm stress test. here's: the result: http://snag.gy/PO519.jpg
As you see Mantle performed worse than dx
Can anyone try this on a same card? I tried other settings (high, low), other scenarios (follow, rts) but i didn't see anything that proved mantle is faster than directx...

Thanks in advance!

 

AMD Catalyst 14.6 Beta drivers are also notable because they bring ... and Mantle is now enabled for notebook models based on AMD Enduro ..

 

The feeling of dissapoint when you find out mantle is not supported on your old faithful friend 6990m.

 

Much seconded.

 

Hey guys,

Could anyone confirm if there is any improvement for crossfire support in this driver?

Thx