The AMD Mantle Thread

[Shivansps]

I read it all a long time ago & it's illogical to bash AMD.

They claim 900% performance increase over DX...

In draw calls

 

[Gloomy]

http://www.golem.de/news/amds-mantle-api-im-test-der-prozessor-katalysator-1402-104261-3.html


Almost 80% faster when fully CPU bound. This is expected for me.

3770K + 290X, 64 players

 

[sontin]

Also the Ti is a 700$ card. So matching its performance at a lower price point is a good thing.

Only in BF4. In AC4 a GTX780TI is 30% faster.

http://www.golem.de/news/amds-mantle-api-im-test-der-prozessor-katalysator-1402-104261-3.html

Almost 80% faster when fully CPU bound. This is expected for me.

On AMD hardware. I wait for tests with an nVidia card.

 

[desprado]

3770K + 290X, 64 players

Better Reviews are from Guru3d and Anandtech i dont believe foreign benchmark

As far i understand is that Mantle performance get much more down in higher Res.

So far at 1080p there is around 1o to 15% improvement which was at least to expect but at 1440p or 1600p 8 to 8% and 4k it is useless.

 

[cytg111]

To be fair, Mantle doesn't really change much for high end, unless you're bottlenecked CPU wise. This will however massively help out buyers who will buy cheap CPUs instead and spend more money on video cards.

Right, how is that not high end? You get to take your high end CPU and add GPU compute power to it until you ARE CPU limited. Crossfire? Trifire? Quadfire? .. You will have to explain highend for me then?

 

[Grooveriding]

Gloomy u have to know that games like BF4 there no difference in 3960x and 3770k.

As what guru3d says
The gains might now be massive, but at enthusiast level these remain to be significant and intersting gains alright. Above is the final example of how I think you guys play the game right now, at this very moment. Obviously 1440P is 2560x1440 WQHD btw and even there you'll see a 10% performance increase.

But 10% free performance is also considered as a massive achievement.

3970X is about 40-50% faster than a 4670K in Battlefield 4.

As far as 10% on top end equipment. That is decent enough.

Anyone see results using a 4670K or 8350 ?

 

[desprado]

Only in BF4. In AC4 a GTX780TI is 30% faster.



On AMD hardware. I wait for tests with an nVidia card.

Bro i tried with single GTX 780 in that campaign where dice claim 58% with CF guess what there was barely 10fps difference with settings ultra,2 Msaa @ 1080p

 

[jj109]

3770K + 290X, 64 players

Either something is seriously wrong with AMD drivers or I've got a golden build of DX11.

 

[desprado]

3970X is about 40-50% faster than a 4670K in Battlefield 4.

As far as 10% on top end equipment. That is decent enough.

Anyone see results using a 4670K or 8350 ?

Difference can be done by a bit of overclocking.Still there is no difference.

 

[mikk]

Better Reviews are from Guru3d and Anandtech i dont believe foreign benchmark

These tests are far worse since nobody tested fully CPU bound and GPU bound in full player MP so far. Marc Sauter is one of the few who knows how to test CPUs/GPUs properly (former PCGH tester). Not to mention some noobs tested against Windows 7.

 

[Grooveriding]

Difference can be done by a bit of overclocking.Still there is no difference.

A bit of overclocking on a 4670K will make up the 40-50% difference ? Of course we can't overclock the 3970X I guess

Mantle could make a huge difference in Battlefield 4 multiplayer for anyone running a 2500K, 3570K or 4670K, which is quite a few people. I'd wait to see more fully fleshed out benchmarks.

Even 10% with a 3970X & 290X is somewhat impressive, nothing amazing, but it's not negligible. That is the sort of performance difference you see posters on the forum refer to as 'crushed' when it's a lead for their favourite brand.

 

[Skurge]

3770K + 290X, 64 players

Well thats a slam dunk right there. 3770K is a high end CPU. 290X is a high end GPU. The minimums in Mantle are higher than the Avg in DX. How is this not awesome?

 

[Shivansps]

3770K + 290X, 64 players

So it only lost 4 fps by going from 720 to 1080P? (so cpu limited at 720P?) them it lost around 60% FPS by using 4xMSAA?!

 

[Gloomy]

Either something is seriously wrong with AMD drivers or I've got a golden build of DX11.

Unless something has changed since [H] did their BF4 performance review (game was buggy back then so I wouldn't count it out) both AMD and Nvidia have similar enough performance.

They've revisited BF4 MP in several video card reviews since then and haven't reported anything out of the ordinary with AMD cards.

So if something seems out of line, then it's probably the review, not AMD drivers. But then again you never know. Maybe the processor is at stock.

 

[DamnedLife]

Don't know German so can't say if that guy tested SP or MP. I'd like to see how that 3960X would compare if they had underclocked it and/or shut off some cores.

He says SP, since MP with full 64 players servers are not repeatbale... For me runinng once would be enough though, give numbers in a great match with lots of explosions in shangai and I'm content.

 

[wand3r3r]

A bit of overclocking on a 4670K will make up the 40-50% difference ? Of course we can't overclock the 3970X I guess

Mantle could make a huge difference in Battlefield 4 multiplayer for anyone running a 2500K, 3570K or 4670K, which is quite a few people. I'd wait to see more fully fleshed out benchmarks.

Even 10% with a 3970X & 290X is somewhat impressive, nothing amazing, but it's not negligible. That is the sort of performance difference you see posters on the forum refer to as crushed when it's a lead for their favourite brand.

Exactly.

I'm still waiting on the 290x crossfire results.

In the end it's interesting and I didn't buy my gpus for mantle, it's just a fringe "benefit" of a yet to be seen amount.

The amount of goal post shifting, while not surprising is sad.

 

[krumme]

http://www.golem.de/news/amds-mantle-api-im-test-der-prozessor-katalysator-1402-104261-3.html

WOW. In MP mantle kills DX. KILLS IT.

 

[Arkadrel]

3770K + 290X, 64 players

1920x1080 and a ~75% increase in performance.

Lmao that seems like quite the jump there.

I guess mantle is okay for people with 3770k that dont overclock their CPUs.

So it only lost 4 fps by going from 720 to 1080P? (so cpu limited at 720P?) them it lost around 60% FPS by using 4xMSAA?!

Early beta mantle driver doesnt like MSAA, it looks like.

 

[jpiniero]

He says SP, since MP with full 64 players servers are not repeatbale... For me runinng once would be enough though, give numbers in a great match with lots of explosions in shangai and I'm content.

OK, that makes sense. Sounds like SP isn't demanding enough to make much of a difference.

 

[Gloomy]

1920x1080 and a ~75% increase in performance.

Lmao that seems like quite the jump there.

I guess mantle is okay for people with 3770k that dont overclock their CPUs.

Hit me up when you get a 75% overclock on a 3770K

 

[mikk]

I wonder how big the gain is on a 8 core CPU like FX-8350. Considering that Mantle should scale better with cores the gain should be bigger there compared to an Intel 4/8 Quadcore, maybe a doubling is theoretically possible, even though in practical game settings with MSAA the gains are much smaller.

 

[cytg111]

Right, how is that not high end? You get to take your high end CPU and add GPU compute power to it until you ARE CPU limited. Crossfire? Trifire? Quadfire? .. You will have to explain highend for me then?

- Aaaand:

http://www.guru3d.com/articles_pages/amd_mantle_preview,2.html

"On massively CPU limited situations like low/medium quality settings / low resolutions or even Multi-GPU setup, that's where the gain easily can get 30 even 40%. As such Mantle will benefit low-end and mainstream processor based PCs the most. But even at Enthusiast class, it will make a difference."

Lets dumb it down :

" low/medium quality settings / low resolutions or even Multi-GPU setup, that's where the gain easily can get 30 even 40%"

"easily can get 30 even 40%"

"easily"

"Multi-GPU setup"

"Enthusiast"

Ill be looking forward to reading the troll-spin that will turn that into a massive fail. Have a go at it guys.

 

[sontin]

As long as there is no comparision to nVidia it's so hilarious to read the hype material.

 

[krumme]

BANG.
Slam dunk for bf4. That is damn great news. Everyone is happy ! Now give me my driver

 

[cytg111]

1920x1080 and a ~75% increase in performance.

Lmao that seems like quite the jump there.

I guess mantle is okay for people with 3770k that dont overclock their CPUs.


Early beta mantle driver doesnt like MSAA, it looks like.

Strong first showing.. need to complete the picture though.