The AMD Mantle Thread

[SiliconWars]

Thief Mantle slides - http://i.imgur.com/nt5217p.png

Some interesting stuff regarding smoothness.

Also

PC Gamer will get higher framerate in GPU limited cases.

 

[monstercameron]

The Iris Pro is being used in the macbook pro portable...the most popular high end portable on the current market and the biggest seller (for high end).The macbook pro with Iris Pro has over 10 hours of battery life:



http://www.theverge.com/2013/10/30/5044874/13-inch-macbook-pro-with-retina-display-review-2013

The 15 inch macbook pro with Iris Pro lasted just under 10 hours of battery life:

http://www.theverge.com/2013/10/24/5024346/macbook-pro-with-retina-display-review-15-inch-2013

Endgadget tested over 11 hours of battery life with IRis Pro:

http://www.engadget.com/2013/10/29/macbook-pro-with-retina-display-review-13-inch-2013/

You brought the 30 minutes of Iris Pro battery life up. I guess that isn't the case. Look, I get the argument that Mantle will help raise the performance of AMD APUs. I am curious as to which AMD mobile APU has 10 hours of battery life, though, and then we can get a fair judgement of what the performance increase is come December. Do you feel that this will help propel an A4-1200 past Iris Pro in BF4? Just curious. Like I said, i'll reserve final judgement till then. I don't know. Maybe Mantle is the most amazing thing ever - I expect a big performance jump but I don't think it's realistic to expect miracles. Heck, if Mantle is the most amazing thing ever i'll be the first to praise it. We'll see.

I wasn't mentioning mantle, just that you called the iris pro toting i7's "mobile", when they can use as much power as a desktop...also what does an a4-1200 what to do with anything? don't reply to me with such silly rhetoric...please!

look under load this will eat over 50W and this isnt even the top spec i7
http://www.notebookcheck.net/Review-Apple-MacBook-Pro-Retina-13-Late-2013-Notebook.105035.0.html

top spec will eat over 70W, how are these classified to even be "mobile"
http://www.notebookcheck.net/Review-Schenker-S413-Clevo-W740SU-Notebook.98313.0.html

 

[Saylick]

Thief Mantle slides - http://i.imgur.com/nt5217p.png

Some interesting stuff regarding smoothness.

Interesting, indeed.

So from that link, the key points, in my opinion are:

- Mantle still uses an abstraction layer, albeit this layer is thinner than the layer used in DX. This new layer allows developers to utilize the CPU and GPU more efficiently. This new thin layer, combined with the extra functionality afforded to developers, is where the performance gains come from.

- Mantle was built to take advantage of certain key features of the GCN architecture, but any architecture which supports a minimum feature set will be able to run Mantle. Again, the point is that Mantle is not vendor specific. These features aren't unique to AMD, so if nVidia builds an architecture which has similar features to GCN, it will be capable of running Mantle. However, because next-gen consoles are GCN-based, it makes sense that Mantle caters to the GCN architecture and will thus support GCN out-of-the-box.

-A lot of the optimization effort falls back on the developer's end, but this is a good thing because we want developers back into the driver's seat. More developer control means better optimizations, which means better games. This extra "burden" on developers isn't much different than the optimization process they use for extracting as much performance out of consoles, so this isn't new for them. In other words, developers are very familiar with this process. Note: Mantle is NOT used on the consoles, but again, the optimization process is similar so transitioning to Mantle from consoles feels natural.

-As a result, you get better CPU and GPU utilization, i.e. you get better performance using the exact same hardware. Without Mantle, having only 50% of the GPU being utilized was not uncommon.

-Native DX ports will not get much better performance because the render pipeline is unchanged. In order to extract better performance, the render pipeline must be rewritten to be more efficient; Mantle allows the developer to do exactly this. Therefore, the render pipeline must be rewritten differently instead of "emulating" DX.

 

[OCGuy]

Thief Mantle slides - http://i.imgur.com/nt5217p.png

Some interesting stuff regarding smoothness.

Also

Wait a minute....the PC gamer might get better performance in GPU limited cases? I swear when Mantle was announced, it was to be the death of all things M$ and NV. Someone even said it may be comparable to what a very popular Apple product did for that company a few years ago.

I don't know what to believe anymore....!

 

[SiliconWars]

The point was, plenty of people are claiming it's just all about removing the CPU bottlenecks and that was basically where all the performance improvements were coming from.

 

[desprado]

The point was, plenty of people are claiming it's just all about removing the CPU bottlenecks and that was basically where all the performance improvements were coming from.

Yes it will improve performance but not that much that people hope ad hyped for.

 

[SiliconWars]

Yes it will improve performance but not that much that people hope ad hyped for.

The numbers being talked about are 20-50%, which is pretty much what most people (at least the AMD guys) said it would be.

 

[desprado]

The numbers being talked about are 20-50%, which is pretty much what most people (at least the AMD guys) said it would be.

AMD guys said nothing and if they know how much performance gain they would tell by now.

See this
As reported by tech report:
Andersson didn't bring up performance estimates, but other developers who discussed Mantle at APU13 did. Jorjen Katsman of Nixxes, the firm porting Thief to the PC, mentioned a reduction in API overhead from 40% with DirectX 11 to around 8% with Mantle. He added that it's "not unrealistic that you'd get 20% additional GPU performance" with Mantle.

Guys say u could get 20% performance although he did not mean at least so what mean to say he not sure but 20% is there target.

 

[SiliconWars]

Yes and the only demo we've seen so far has Mantle 3 times faster.

 

[Silverforce11]

Yes and the only demo we've seen so far has Mantle 3 times faster.

Only in extreme cases where the CPU is the bottleneck, ie. massive AI fights, strategy games, MMOs are the likely cases where massive gains are possible. Traditional FPS where the GPU is the bottleneck, I would say 20% is a decent outcome.

What's interesting in the Frostbite presentation, is that they claimed Mantle offers a more efficient approach to handle AA. There's a lot of unknown, but I tend to agree with statements that single digit performance gains would be "meh" and not deserving of years of hard work to get Mantle going. It needs to offer a huge performance gain.

 

[Saylick]

Yes and the only demo we've seen so far has Mantle 3 times faster.

I doubt you'd see 3x faster for more than 5% of the total rendering time, but this is a good thing nonetheless. Not only does it showcase the potential of Mantle, it also hints that FPS minimums might come up significantly. Of those times where performance was improved by 3x, the game was most likely 100% CPU bottlenecked and thus most likely correlate with the lowest FPS readings out of the entire run. With Mantle, those bottlenecks disappear and minimums get a huge uplift.

In reality, we should expect a lot less uplift on average. 20-50% seems like a good guesstimate. :thumbsup:

 

[desprado]

Yes and the only demo we've seen so far has Mantle 3 times faster.

ur answer is here that is why i say hype but dont hype to much

This is what Mantle does. It bypasses DirectX (and possibly the hardware abstraction layer) and developers can program very close to the metal with very little overhead from software. This lowers memory and CPU usage, it decreases latency, and because there are fewer “moving parts” AMD claims that they can do 9x the draw calls with Mantle as compared to DirectX. This is a significant boost in overall efficiency. Before everyone gets too excited, we will not see a 9x improvement in overall performance with every application. A single HD 7790 running in Mantle is not going to power 3 x 1080P monitors in Eyefinity faster than a HD 7970 or GTX 780 (in Surround) running in DirectX. Mantle shifts the bottleneck elsewhere.
http://www.pcper.com/category/tags/dx11

Mostly there target is 20% as dev say in my recent post

 

[GaiaHunter]

Wait a minute....the PC gamer might get better performance in GPU limited cases? I swear when Mantle was announced, it was to be the death of all things M$ and NV. Someone even said it may be comparable to what a very popular Apple product did for that company a few years ago.

I don't know what to believe anymore....!

What I read is "PC gamer will get higher frame rate in GPU limited cases".

That "might" sounds like wishful thinking.

 

[GaiaHunter]

Mostly there target is 20% as dev say in my recent post

20% on a new tool with a few months of use instead years and years of use?

Not bad.

Wait until they actually know what they are doing.

 

[SiliconWars]

Games are GPU and CPU bound at various stages. Mantle will remove the CPU limitation in nearly all cases. Obviously the bigger gains are with CPU bound scenarios, but the latest slides show that the gains are still large in GPU bound scenarios as well.

20% is low-balling, and this doesn't factor in that the IQ will also be improved. Very few devs are going to use Mantle simply for fps improvements, most will be more interested in adding IQ. BF4 on Mantle could end up with double the draw distance and still be 30% faster. However you look at it, it's really bad news for Nvidia.

 

[desprado]

Only in extreme cases where the CPU is the bottleneck, ie. massive AI fights, strategy games, MMOs are the likely cases where massive gains are possible. Traditional FPS where the GPU is the bottleneck, I would say 20% is a decent outcome.

What's interesting in the Frostbite presentation, is that they claimed Mantle offers a more efficient approach to handle AA. There's a lot of unknown, but I tend to agree with statements that single digit performance gains would be "meh" and not deserving of years of hard work to get Mantle going. It needs to offer a huge performance gain.

I will say no to that.Even Dev has said that 20% performance gain is the target usually and if more that is the advantage.
first of all it is an API and it is not magic that 50 to 70% performance just from API.

 

[desprado]

Games are GPU and CPU bound at various stages. Mantle will remove the CPU limitation in nearly all cases. Obviously the bigger gains are with CPU bound scenarios, but the latest slides show that the gains are still large in GPU bound scenarios as well.

20% is low-balling, and this doesn't factor in that the IQ will also be improved. Very few devs are going to use Mantle simply for fps improvements, most will be more interested in adding IQ. BF4 on Mantle could end up with double the draw distance and still be 30% faster. However you look at it, it's really bad news for Nvidia.

That i agree even if Mantle games are 20% at most that is really bad sign for Nvidia and that cloud give a wake up call.

 

[SiliconWars]

I think the real problem is it'll be 20% at worst. There is a wide variety of games - maybe most FPS types will only see 20% gains, but all the RTS and strategy games could and should be a lot more.

Even if some devs just do Mantle version FPS with increased draw distance and nothing else, it's still going to be a massive selling point for AMD.

It actually just gets better the more you think about it. With asymmetric crossfire and the combined pipeline, APU's just add even more performance to the AMD cards. We'll see plenty of situations where Kaveri and a 290X will easily beat the 4770K and 780 Ti, simply because of the extra shading power of Kaveri.

Even if Nvidia licenses Mantle or AMD just lets them in on it for free, they'll never be able to get over this deficit of combined APU + discrete. In a way AMD can't lose even if they just open up Mantle to Nvidia freely - industry adoption would then be total, AMD's APU's would have hugely increased value to gamers and Nvidia would still be behind. This is way, way bigger than most people still realise.

 

[OCGuy]

Games are GPU and CPU bound at various stages. Mantle will remove the CPU limitation in nearly all cases. Obviously the bigger gains are with CPU bound scenarios, but the latest slides show that the gains are still large in GPU bound scenarios as well.

20% is low-balling, and this doesn't factor in that the IQ will also be improved. Very few devs are going to use Mantle simply for fps improvements, most will be more interested in adding IQ. BF4 on Mantle could end up with double the draw distance and still be 30% faster. However you look at it, it's really bad news for Nvidia.

You understand this is the kind if hyperbole that hurts AMD on the PR front I hope. Professional PR people are paid to create slides and set industry expectations, which at any company are generally optimistic in nature.

Then someone reads this, and holds what could possibly be a successful product launch to an impossible standard. Do you have any thing solid to support your numbers?

 

[SiliconWars]

You understand this is the kind if hyperbole that hurts AMD on the PR front I hope. Professional PR people are paid to create slides and set industry expectations, which at any company are generally optimistic in nature.

Then someone reads this, and holds what could possibly be a successful product launch to an impossible standard. Do you have any thing solid to support your numbers?

Yes, various sources have said 20%-50% improvements. There is a video with 3x higher performance. The BF4 chief architect has said Mantle will have better performance and better graphics.

What more do you want?

 

[desprado]

Yes, various sources have said 20%-50% improvements. There is a video with 3x higher performance. The BF4 chief architect has said Mantle will have better performance and better graphics.

What more do you want?

Post ur source which says 50% gain performance or else edit ur post plz.What do mean by better graphics.It is not a graphic engine it is just a API.So i guess u want to say PS4 API has better graphics?

 

[SiliconWars]

http://fudzilla.com/home/item/33137-amd-sheds-more-light-on-mantle

So how much of a performance boost are we looking at? Well AMD claims 20 to 50 percent, which sounds very impressive indeed

http://i.imgur.com/nt5217p.png

More expensive compute-based effects possible with a low performance impact

You can render a lot more

...or games can get more complicated without losing performance - Eg increased draw distance

It's all there if you just want to look.

 

[OCGuy]

Yes, various sources have said 20%-50% improvements. There is a video with 3x higher performance. The BF4 chief architect has said Mantle will have better performance and better graphics.

What more do you want?

Links and direct quotes make credible discussion possible. Obviously you can do what you want, but...

 

[desprado]

http://fudzilla.com/home/item/33137-amd-sheds-more-light-on-mantle



http://i.imgur.com/nt5217p.png







It's all there if you just want to look.

If u read more carefully he said that
so we're probably looking at something closer to 20 percent
but still it will hurt Nvidia until they figure out something

But that performance will not hurt a MS a bit or nether DX and it will only hurt Nvidia

 

[OCGuy]

Ah ok.....when you give percentages, I tend to think FPS. "Performance" is a weasel word and cannot be measured.