Oxide Games Says AMD Mantle is Disruptive Technology

[exfileme]

Dan Baker talks about using Mantle for Star Swarm.

Oxide Games Says AMD Mantle is Disruptive Technology : Read more

 

[heero yuy]

misleading title much?he says its the (disruptive) technology that would get everyone to rethink thingsthe way the title puts it is more to the effect that he said its a bad technology - the complete opposite as to what he actually said

 

[nickul]

...and then, somewhere, at the very end, in the shortest statement ever: "Later, he said that Mantle is somewhat of a disruptive technology in that it is getting everyone into rethinking things."What a nice way of minimizing all the good things said about Mantle in that interview, and maximizing that one single aspect that is only somewhat negative => Nice chosen title...<= Niice... Nice AMD bashing... again.... I see you're keeping up the good work, Tom's Hardware! U know what, I volunteer to help you start a rumor about how disruptive Mantle is!I will post on Facebook! I'll tweet! I will paint my cat and blame it on Mantle!! XD

 

[chicofehr]

Mantle main problem is its going to take some time for developers to optimize their own software to utilize it to its full potential. Most developers have been using DX for very long and using a new api takes time to learn. I am sure the performance will be way better once they get a hang of it.

 

[mstngs351]

It's funny how many people are butt hurt over the use of the word "disruptive". It doesn't actually have any negative connotation unless you assign it. Disruptive: to cause (something) to be unable to continue in the normal way

 

[Soda-88]

Kevin Parrish, you should write for The Sun or Daily Mail.

 

[dotaloc]

I hate misleading titles and this one doesn't even bother me. He could have also reasonably titled: "Mantle Dwarfs D3D Performance" or "Mantle: The Elegant Monster Stirs."

 

[gwwerner]

For all of you who are upset about the word disruptive, from wikipedia:A disruptive innovation is an innovation that helps create a new market and value network, and eventually disrupts an existing market and value network (over a few years or decades), displacing an earlier technology. The term is used in business and technology literature to describe innovations that improve a product or service in ways that the market does not expect.

 

[Kewlx25]

Mantle main problem is its going to take some time for developers to optimize their own software to utilize it to its full potential. Most developers have been using DX for very long and using a new api takes time to learn. I am sure the performance will be way better once they get a hang of it.

They said that a crappy implementation of Mantle code still kicks the crap out of well optimized DX11 code. These people specialize in DX11 and they can slap together random crap in Mantle and make it work better.

They also mentioned that when they try to thread their code too much, the driver devs tell them to back off because they're choking the driver threads. Yet the CPUs are mostly idle.

The whole DX/OpenGL paradigm is so bad for modern computers that driver engineers are telling programmers to use less CPU when a CPU is almost entirely idle. The system is broken, it needs to be replaced.

 

[maxiim]

If you look at steam for example, 52.01% of users are running nVidia cards, amd comes in a far second with 31.36%, so if developers were to start using mantle only to cater games to AMD gpu/cpus they would be excluding a at least HALF of the PC market with their games. Mantle is a good idea, but now more than ever it seems more a marketing push for amd cpu/gpus. I just dont think mantle will be as utilized as something like gsync will.

 

[mstngs351]

So Oxide, who was paid to promote Mantle on a tech demo says it didn't put much effort into it. I'll take that with a grain of salt...

 

[airborne11b]

Mantle is something that is a waste of time for anyone but the largest AAA companies. Taking time and resources out to learn a new API and implement it, when it will only benefit 30% of the people buying your game?Complete trash. As a programmer, you have to willing to constantly learn new technologies to succeed, but this is simply terrible. Sure, huge companies like EA might be able to throw a ton of $$$ at silly things like this, but it's not viable for most developers.

 

[MANOFKRYPTONAK]

Toms has changed people are moronic and not helpful. The title is well constructed, it is not the authors fault people don't know the correct meaning or usage of a word... And swearing indicates very low intelligence IMHO.

 

[nitrium]

Yep, SSDs were also considered a "disruptive technology" - because no amount of HDD technology could ever hope to match it, thus "disrupting" the way we look at storage performance. Some of posters here are really not the sharpest knives in the block.

 

[damonese]

About disruptive buzz.. Yeah i got you. Everyone is happy now.. Haha

 

[hotroderx]

I think Mantle is a cool idea and it does sound like its more powerful and a solid replacement for DX. I think the only problem is it will be a gimmick unless AMD opens it to Nvidia and Intel. What incentive does a game manufacture really have to program a game with Mantle on top of DX? I would think AMD would need over 50% of the video card market to make it worth while for game developers. I seriously doubt they have that kinda market saturation at this point. I mean you could argue that AMD's new cards are in such demand there ranking in a premium but at the same time 90% of those cards are being solid to miners. I dont know cool tech but not sure it will make it in this market.

 

[Onihikage]

Somehow most of the people commenting here don't seem to realize that Mantle is mostly on the CPU side, massively boosting performance on systems where the CPU is the bottleneck while only slightly boosting performance on systems that are maxing out the GPU. Regardless of either scenario, ALL systems show some benefit from Mantle, not just those with AMD hardware. The amount they benefit compared to D3D varies, but it's always at least a small benefit in performance. There's really no reason not to use it.

 

[Puiucs]

why the hell is a stupid title that looks like it belongs in tabloids in a tech related news website? are we idiots or something?

 

[childofthekorn]

I'm still curious why the hate over the rebadge. Nvidia does the same thing, I'm still impressed that the general consumer is letting both companies get "away with it" and equally impressed that GFX cards are sustaining these days. You won't see the true innovation in what either companies cards can handle until 4k/8k resolutions become a standard, you get people buying more 3x or 6x setups, then the real horse power produced from either red or green side.

 

[jk47_99]

I want to see Mantle on SteamOS, then building a PC at console prices which can play AAA games at 60fps at 1080p really becomes a reality.

 

[mstngs351]

Mantle main problem is its going to take some time for developers to optimize their own software to utilize it to its full potential. Most developers have been using DX for very long and using a new api takes time to learn. I am sure the performance will be way better once they get a hang of it.

They said that a crappy implementation of Mantle code still kicks the crap out of well optimized DX11 code. These people specialize in DX11 and they can slap together random crap in Mantle and make it work better.They also mentioned that when they try to thread their code too much, the driver devs tell them to back off because they're choking the driver threads. Yet the CPUs are mostly idle.The whole DX/OpenGL paradigm is so bad for modern computers that driver engineers are telling programmers to use less CPU when a CPU is almost entirely idle. The system is broken, it needs to be replaced.

No they said they "We have spent a huge amount of time optimizing around D3D..." In other words he was saying that they are far more accustomed to D3D and basically have learned the ropes using it. He didn't say that they had optimized the code for the demo and considering they were likely paid by AMD... grain of salt. I mean I fully expect it to work better but I'm sure there is some fluff here.

 

[Jaroslav Jandek]

I don't trust Oxide anymore and here's why: in their first Star Swarm Mantle performance numbers, they published the number of draw calls and FPS increase (it was 300% IIRC). The problem is (based on the number of draw calls) that they did NOT use instancing (which is a mistake no senior DX developer makes). With instancing (using the DrawIndexedPrimitive method), you can render many visuals (~1 million for GTX 285) in a single draw call! So one ship type is 1 draw call, guns is another, projectiles is another, etc. (-> you only need a few draw calls to render the Star Swarm demo). If you are curious about this feature, you can download Visual Game Studio and an Instancing demo from the XNA website (if it is still available) and run it yourself (it is basically just download and run - even amateurs should have no issues).

Also, Dan Baker says in the interview: "...D3D made an attempt to be threaded, and failed...". Are you kidding me? My rendering engine can work multi-threaded on as many cores as I want, because ID3D11Device is thread-safe and ID3D11DeviceContext requires simple synchronization (although some drivers do not support D3D11_FEATURE_DATA_THREADING). Sure, it could be better, but it is getting there.

That being said, I am sure Mantle is a better API, since it is newer and is not limited by legacy considerations. The issue is, it may not get implemented by Intel and nVidia and I am personally not going to waste time to target Mantle if it isn't implemented by all the major GPU manufacturers.

 

[vaughn2k]

Gotta love reading these comments, more entertaining than the article itself...

 

[renz496]

in my opinion both intel and nvidia will only adopt Mantle if AMD invites both party to co-develop the API and let the other two have their hands in the API direction in the future. simply open the software for other to use will not let them jump on it.

 

[jin_mtvt]

Are you joking? The only way Intel and Nvidia would consider working with Mantle would be if many large publishers would ask/force them to do so. Jaroslav : are you daydreaming?? If no one suppors mantle because it involves additional work, it won't caught up. No other new API will never unless some labor is put into it. Currently mantle is in its infancy, lets wait to see how it develops before trying to kill it please.There is only possible gain from consumer POV, no one is forcing anybody to buy a GPU that would not work unless using this new API, no risks other than $$ from developers is involved for now. From benches, seems to be a very important performance improvement in low to mid gaming rigs ( which repreents what..95% of the market?? ). Everything about it is still Beta, lets hope they can refine it more.