Oxide has tested their engine on 6 cores and 12 threads (an Intel CPU) in Mantle to see how well it scales.
So could you please stop with the "Mantle only works on AMD CPU's" talk?
I am thinking that if AMD say that Mantle can only be used by AMD CPU what will happen than.
Must resist...must resist...
Someone has to be a voice of reason...I mean there are posters talking about thoundsands of percent of increased performance over DirectX...due to the PR pushing, but actual products failing.
But I am glad (some) people have learned the difference between PR and products now :thumbsup:
desprado said:I am thinking that if AMD say that Mantle can only be used by AMD CPU what will happen than.
Oxide has tested their engine on 6 cores and 12 threads (an Intel CPU) in Mantle to see how well it scales.
So could you please stop with the "Mantle only works on AMD CPU's" talk?
He's actually partially right here. While Mantle has been shown working on both Intel and AMD CPUs, currently it only works systems with AMD GPUs. So an Intel APU on it's own won't be able to run Mantle currently.
He's actually partially right here. While Mantle has been shown working on both Intel and AMD CPUs, currently it only works systems with AMD GPUs. So an Intel APU on it's own won't be able to run Mantle currently.
Funny, I seem to remember several people pointing out the Nitrous engine showing 8.5 FPS under DirectX, and 30 FPS under Mantle (12 FPS DirectX/32 FPS Manlte with Intel I7 4770K CPU), something that Dan Baker of Oxide has said is a 'conservative' estimate on AMD's part. Something that he also said that he was able to accomplish by himself in two months.
Lessee... 30 divided by 8.5 equals... 3.529, or about 3.5x the framerate That's a little bigger than thousandths...
Been posted before, but here are the numbers again. See images 13-14 here. Dan Baker of Oxide's comments track quite nicely with these numbers, and the fps count on the video I linked above seems to track with these numbers as well (actually, the FPS numbers in the interview footage linked above are usually much higher than 30fps)
http://imgur.com/a/Oqmuf#12
Nixxes' Katzman (the Thief guy) revealed during APU '13 that “very early figures from Thief” (which is “not fully running on Mantle yet” showed a big reduction in draw call overhead. “Before, we would often see about 40% of the CPU time stuck in the driver, in D3D, or in various threads,” he said. “The early measurements we did, right now we have that down to about a fifth of that.”
So, according to THAT developer, he saw an approximate 80% reduction in CPU time used for drivers, D3D, and various threads, even though it wasn't fully utilizing Mantle yet. That was about two months ago...
I haven't seen ANYTHING that shows a mere thousands of a percent increase with Mantle. Perhaps you'd like to share your sources with the class?
If you don't have verifiable hard numbers of your own to share, you are of course welcome to wait for the Star Swarm Steam release (this month), the BF4 Mantle patch, and Thief (Feb 25th-ish), along with the rest of us.
In the meantime, when THREE game engine developers continue to tell us that they are seeing huge gains with Mantle, well they are putting their own reputations on the line by doing so, and NONE of them work for AMD so I tend to take their thoughts and comments on the subject a little more seriously than AMD's own comments. After all, the game developers are the ones that get to make the stuff work, AMD just provides the framework.
Must resist...must resist...
Someone has to be a voice of reason...I mean there are posters talking about thoundsands of percent of increased performance over DirectX...due to the [REDACTED] pushing, but actual products failing.
But I am glad (some) people have learned the difference between [REDACTED] and products now
Used [REDACTED] so your head wouldn't explode ^^
Yes. But it was only for the cpu part without any gpu optimization.
There is no way a serious mantle implementation, using all the tech decribed as eg. Johan at apu13, only cost 2 man months for an engine like bf4. He never spoke of 2 man-month either. And thats for up to 45% performance.
Lonbjerg is just low level trolling and having fun. Its rethorical questions.
What remains of fact is, we have oxide showing an impressive engine running, a meager 2 month after they were introduced to mantle. Performance on the application level is up 300% for mantle compared to dx, on what is probably a state of the art RTS engine. The engine uses no lead thread for a starter. Thats with or without mante, and its far beyond any engine until now. Its build to be state of the art at dx.
2 man-month to get 300% performance on top of that - using only the cpu part - is stellar.
What stands is, if this is doable in 2 month - if it was 2 or 10 man months - its the best investment a developer can do, and the most easy decision to take at least on the RTS side. Its a no brainer. Its obvious different for non RTS games. You dont get 300% performance in your engine for 20K usd for FPS. lol.
Infraction issued for callout.
-- stahlhart
Remember this huge difference only exist for rts where you are drawcall limited. Dx is severely limited here.
But the difference is probably going to be even higher when they start to know mantle. A difference of 500-1000% is perhaps the most realistic in a year or 2 and even bigger in 3 years if they directly build the engine to take use of eg. the better multicore scaling in mantle.
But those 1000% difference whatever is easily used in a rts game making eg. your armies bigger or with more individual control of the units. I can see this used besides dx.
But the difference is going to big - yes - but thats the nature of it. It is fundamental gamechanger for the rts genre. But we need those new games for it to develop.
I'll ignore the empty PR in your thread and focus on something fromthe real world (the bolded part).
What I wrote was this:
Based on posts like this:
Any doubts/false notions now?
Funny you only quote only one person and funny you left the part out where he is talking about "rts" and it is possible to have more than 300% boost in rts while having really big armys on screen.
Your cannot see the bolded part with 500-1000%?
If the Oxide game engine manages 100k draw calls with Mantle, but only 15k with DX, you already have the 500% performance increase there.
Let's hope they really release the Starswarm demo this month so we can test it.
#MaxDXFPS 10
#MaxMantleFPS 100that is what i am saying whole time.AMD Mantle slides and everything is indicating about CPU bottleneck.It will be if you are CPU limited, and thats what AMD tells in every slide, Mantle could lead intro a FPS increase in (to say something) Sempron 145, but in a I5-4670K? it may not have a single fps over dx, there is not evidence that points out to a GPU perf increase.
Only way to overcome that is to go WAY over what DX can do you it will CPU limited on every cpu.
No.
Drawcalls is one piece of the puzzle of a frame.
You cannot say 10 x drawcalls more = ~10 x more performance.
Well you could, but that would indicate you have no clue.
It's not like drawcalls are some magical "lever" holding FPS back unilaterally.
You could increase drawcall potential from 15K to 100K in a game...and not gain a single frame...but that would be an outlier
More realistically you would see from 1-~10% preformance increase...on ~5% of SKU's in the market.
Excuse me for not popping the champagne
Well, if you have 10 times more drawcalls with the same IQ and same FPS, wouldn't that indicate 10 times more performance.
So now Mantle is for a niche of games on a niche of SKU's?If in the Oxide demo you have 15k draws at 60 fps with DX, and 60 fps with 15k draws with Mantle, you don't see any performance gains. But it's not about how much fps you can get. It's about how many units you can get while having a decent frame rate.
It's not about having more FPS with the same IQ, it's about pushing what the game can do.
If you are making a game that is trying to simulate massive battles, you want to have as many units as possible. If with DX you can only do 2k units on screen, and with Mantle you can do 10k units under the same IQ and FPS, then in that scenario you have 5 times more performance.
So now Mantle is for a niche of games on a niche of SKU's?
Game changer for mobile if you ask me. Performance aside, CPU overhead=power usage.
Imagine AMD 28nm APU having better perf/watt than 14nm Intel (or intel+nv card).
There are so many places where mantle will change the landscape, that I'm not surprised many have hard time grasping it :whiste:
RTS on a mobile device?
Really?
You know why they are litterally no excistant on consoles right?
The mouse/keyboard control you really need..and you think they are suited for mobile? :ninja:
Really?
Maybe someday, someone will invent how to plug a mouse to a laptop. You never know what will future bring!