The AMD Mantle Thread

[Gloomy]

Well, it's no news that Mantle doesn't help much in a GPU limited scenario.

Computerbase saw an 18.6% increase for a i7-4770K OC with R9 290X:
http://translate.google.com/transla...efield-4-mit-mantle-erster-eigener-benchmark/

Much better

Keep the reviews coming...

 

[BrightCandle]

It sure took a lot of effort from AMD and dice to achieve a moderate increase in performance. Its not nothing, but I don't think so far the numbers suggest that every game should spend 2 dev months (well lets be honest its been a lot more than that I suspect) implementing the Mantle interface for 15% extra performance. Kind of needs to be more like 50% in all scenarios not a few scenarios which most people (APUs seriously who games on those with a high end GPU!) don't use.

 

[sontin]

Yeah, i dont see any reason to spend so much time for 10% more performance.
And even then AMD is not really faster than nVidia.

 

[Gloomy]

Yeah, I'm not really seeing the point unless I really get 18% from my 4770K+R9290. 10% is nothing... at 18% maybe I'll notice the difference.

If Dice's Crossfire results pan out then good. But that's a fraction of a fraction of a fraction of the market. Though I guess a number of folks on this forum will get their money's worth.

Overall not really a failure but not worth a 250 page thread full of heartache and bitterness. That's the real tragedy here. It wasn't worth the hype.

 

[blackened23]

~5% gains mostly for high end hardware. Meh. This is not what AMD or their extensive social media marketing team hyped Mantle up to be months ago, IMO. Pretty sure everyone expected higher gains than this, as of right now for high end hardware it is rather meh. It isn't nothing, it's free performance for AMD users which is a good value add for them (assuming you're using an APU or 290/X). But this isn't the Mantle that was hyped so heavily months ago, this isn't what we were led to believe. The hype led us to believe the gains would be unreal, and nobody expected the gains to be primarily for low end bargain bin CPUs. Nobody who buys a 290X for 600-700$ is going to pair it with a cheap junk HTPC chip - they're going to match it with a similarly high end CPU and motherboard.

 

[mikk]

That's actually pretty good all things considering.

Not really. Almost no difference tested against Win 8.1. In GPU limits only minor differences as expected.

 

[desprado]

~5% gains mostly for high end hardware. Meh. This is not what AMD or their extensive social media marketing team hyped Mantle up to be months ago, IMO. Pretty sure everyone expected higher gains than this, as of right now for high end hardware it is rather meh. It isn't nothing, it's free performance for AMD users which is a good value add for them (assuming you're using an APU or 290/X). But this isn't the Mantle that was hyped so heavily months ago, this isn't what we were led to believe. The hype led us to believe the gains would be unreal, and nobody expected the gains to be primarily for low end bargain bin CPUs. Nobody who buys a 290X for 600-700$ is going to pair it with a cheap junk HTPC chip - they're going to match it with a similarly high end CPU and motherboard.

it is not worth much for high end user but mid to low it is a different story.

 

[Crap Daddy]

~5% gains mostly for high end hardware. Meh. This is not what AMD or their extensive social media/forum marketing team hyped Mantle up to be months ago, IMO. Pretty sure everyone expected higher gains than this, as of right now for high end hardware it is rather meh. It isn't nothing, it's free performance for AMD users which is a good value add for them (assuming you're using an APU or 290/X). But this isn't the Mantle that was hyped so heavily months ago, this isn't what we were led to believe. The hype led us to believe the gains would be unreal, and nobody expected the gains to be primarily for low end bargain bin CPUs. Nobody who buys a 290X is going to pair it with a cheap low end HTPC chip - they're going to match it with a similarly high end CPU and motherboard.

It's really nothing new coming from AMD. I still remember the bulldozer hype. Now that was the real deal as far as hyping. At least mantle doesn't perform worse than dx.

 

[blackened23]

it is not worth much for high end user but mid to low it is a different story.

True, but there are 2 confirmed Mantle games (Thief, BF4) for 2014 and 2-3 "maybe" games. Unless you limit yourself to playing those games only there is no benefit, and PC gamers generally play a lot of games - not just 1 game. Especially since PC games are very cheap with Steam sales.

I don't know. IMO, this isn't the narrative that AMD hyped up months ago at GPU14. We were led to believe the gains would be incredible for high end hardware, because their Mantle marketing was tied in with the high end 290 and 290X. At no point did they give a disclaimer stating "FOR LOW END CPUs ONLY".

That said, it's free performance and it is a value add for AMD GPU users. I just don't think the actual results match the pre-marketing and pre release hype that AMD fed us. Not even close to matching the hype. Which isn't surprising given AMD's marketing history, but still - like I said, it IS a gain for AMD users which is good for them.

 

[sontin]

it is not worth much for high end user but mid to low it is a different story.

Low to mid range will not run into a CPU limit. So Mantle doesn't help them in anyway to improve their experience.

There is no market for Mantle on the PC. In the console world it makes because they combine low performance cores with a powerful GPU. But something like this doesn't exist here and the increase is around 10%.

And this is the same range nVidia and AMD has been archived through their drivers.

 

[desprado]

Yeah, I'm not really seeing the point unless I really get 18% from my 4770K+R9290. 10% is nothing... at 18% maybe I'll notice the difference.

If Dice's Crossfire results pan out then good. But that's a fraction of a fraction of a fraction of the market. Though I guess a number of folks on this forum will get their money's worth.

Overall not really a failure but not worth a 250 page thread full of heartache and bitterness. That's the real tragedy here. It wasn't worth the hype.

This...............
Now it is conformed Mantle will have very Limited games.
I am sure after these results non dev will be willing to invest and it will stay with 2 or 3 dev to use it.

 

[ICDP]

~5% gains mostly for high end hardware. Meh. This is not what AMD or their extensive social media marketing team hyped Mantle up to be months ago, IMO. Pretty sure everyone expected higher gains than this, as of right now for high end hardware it is rather meh. It isn't nothing, it's free performance for AMD users which is a good value add for them (assuming you're using an APU or 290/X). But this isn't the Mantle that was hyped so heavily months ago, this isn't what we were led to believe. The hype led us to believe the gains would be unreal, and nobody expected the gains to be primarily for low end bargain bin CPUs. Nobody who buys a 290X for 600-700$ is going to pair it with a cheap junk HTPC chip - they're going to match it with a similarly high end CPU and motherboard.

Overclocked high end CPU with high end GPU = 18% performance gain.
http://translate.google.com/transla...efield-4-mit-mantle-erster-eigener-benchmark/

So a very good increase in that test. Or should it be ignored as it doesn't suit your agenda?

That chart you posted is NOT ~5% gain, it shows ~10% gain at 19290x1080 and ~8% gain at 2560x1600. ~5% gains at 4K but that doesn't matter remember?

Funny how 4K is relevant all of a sudden.

Not to mention this is still in beta and the performance will improve from AMD over time.

 

[Gloomy]

I want to see results for the 7970. Do we have any yet? Just to see if the current driver supports them even a little.

 

[DarkKnightDude]

Any Min fps numbers?

 

[Arkadrel]

@blackened23

~5% gains mostly for high end hardware. Meh. This is not what AMD or their extensive social media/forum marketing team hyped Mantle up to be months ago, IMO.

Early beta driver, things will probably get abit better.

also from that picture bench you posted:

1920x1080 1xAA 1xAF : 111,9 fps / 102,6 fps * 100% = 109%
Mantle gave a 9% increase (not 5%)

1920x1080 4xMSAA 16xAF : 78,9 fps / 70,8 fps * 100% = 111,5%
mantle gave a 11,5% increase (not 5%)

Even at 2560x1600 resolution its 8%-9%.
At 4k resolution the differnce is only ~7% but fps is to low to be playable.


This test was done with a:
Intel Core i7-3960X @ 3.9ghz
AMD Radeon R9 290X (stock)


Maybe people with slower CPUs, that have a 290x will benefit more?

 

[richaron]

Lol people pay $hundreds in hardware for a few % performance & some of you guys are whinging about it for free.

Hundreds of dollars worth of hardware with a free update and you still need to bash AMD. Seems stupid to me.

 

[DarkKnightDude]

Lol people pay $hundreds in hardware for a few % performance & some of you guys are whinging about it for free.

Hundreds of dollars worth of hardware with a free update and you still need to bash AMD. Seems stupid to me.

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.

 

[DiogoDX]

Not a single multiplayer test yet. So much for the 24 hours press exclusive.

 

[jpiniero]

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.

 

[Arkadrel]

Not a single multiplayer test yet. So much for the 24 hours press exclusive.

Multiplayer should be more CPU intense and show off mantle in a slightly better light.

But you cannot really replicate a multplayer benchmark to 100% simular/fair run thoughs time after time.

Which is why they are all doing single player runs, I assume.

 

[sontin]

Overclocked high end CPU with high end GPU = 18% performance gain.
http://translate.google.com/transla...efield-4-mit-mantle-erster-eigener-benchmark/

So a very good increase in that test. Or should it be ignored as it doesn't suit your agenda?

That chart you posted is NOT ~5% gain, it shows ~10% gain at 19290x1080 and ~8% gain at 2560x1600. ~5% gains at 4K but that doesn't matter remember?

Funny how 4K is relevant all of a sudden.

Not to mention this is still in beta and the performance will improve from AMD over time.

You know that in the same computerbase benchmark the GTX780Ti archives 63,8 FPS with an older BF4 version and driver?

So even with Mantle AMD is only able to catch the GTX780TI. And in higher resolutions they could even be slower.

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.

Single player.

 

[desprado]

Lol people pay $hundreds in hardware for a few % performance & some of you guys are whinging about it for free.

Hundreds of dollars worth of hardware with a free update and you still need to bash AMD. Seems stupid to me.

AMD did a good job.The bashing is for hyping.Read AMD old statement about mantle around Sep to Oct than u will now why people are bashing it.

 

[mikk]

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.

 

[rtsurfer]

You know that in the same computerbase benchmark the GTX780Ti archives 63,8 FPS with an older BF4 version and driver?

So even with Mantle AMD is only able to catch the GTX780TI. And in higher resolutions they could even be slower.

Everyone knows that R9 series cards are almost on par if not faster than Nvidia cards at higher resolution.

Benifiting from AMD's wider Bus.

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

 

[richaron]

AMD did a good job.The bashing is for hyping.Read AMD old statement about mantle around Sep to Oct than u will now why people are bashing it.

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