Home / Professional / Development / AMD: Vulkan absorbed ‘best and brightest’ parts of Mantle

AMD: Vulkan absorbed ‘best and brightest’ parts of Mantle

Early this week Advanced Micro Devices advised game developers to use Khronos Group’s Vulkan and Microsoft Corp.’s DirectX 12 application programming interfaces instead of its own Mantle API. Later on, the chip developer made further clarification: Vulkan has absorbed the key features of Mantle, but it is not proprietary. While this explains why AMD praises Vulcan, unfortunately, this does not reveal the fate of Mantle.

ATI Technologies and then Advanced Micro Devices have always relied on open-standard technologies as major enablers for their hardware. Unlike Intel Corp. and Nvidia Corp., neither ATI nor AMD have invested in proprietary tools, middleware or closed platforms in a bid to improve performance or sales of their products. As a result, when AMD introduced Mantle in 2013, it came as a major surprise. Nonetheless, it was a logical thing to do. Modern graphics processing units are incredibly complex and contain a lot of hidden horsepower that cannot be accessed using traditional APIs. Moreover, modern computer graphics sub-systems at large (which include GPUs, drivers, software, APIs, operating system and so on ) are not really efficient. As a result, AMD needed an API which could exploit its GCN [graphics core next] architecture and would boost performance of Radeon hardware compared to Nvidia GeForce graphics cards in select games.

amd_radeon_hd_7000_artwork

The key targets for AMD when it comes to Mantle were higher framerates, reduced rendering latency, reduced GPU power consumption, better use of multi-core CPUs, and re-pioneering new features like split-frame rendering. Many of those targets have been achieved. A number of games that use Mantle give advantages to AMD Radeon graphics adapters.

However, a proprietary API costs money to develop and maintain. If there is an open-standard cross-platform API that can provide AMD the same advantages as Mantle, then the latter is simply not something that money-starving AMD should invest in. Apparently, AMD worked closely with Khronos and many parts of Mantle became parts of Vulkan.

vulcan_logo

“The cross-vendor Khronos Group has chosen the best and brightest parts of Mantle to serve as the foundation for Vulkan,” said Robert Hallock, the head of global technical marketing at AMD.

The Vulkan is a new low-overhead/high-throughput API that provides access to graphics and compute on modern graphics processing units used in various devices. Just like AMD Mantle, the open-source Vulkan API will enable explicit GPU control, will minimize driver overhead and will enable efficient CPU multi-threading. Since Vulkan will eventually be found on virtually all types of hardware, it makes a great sense for AMD to focus on this API.

“Vulkan paves the way for a renaissance in cross-platform and cross-vendor PC games with exceptional performance, image quality and features,” stressed Mr. Hallock.

amd_mantle

Officially, AMD does not want to proclaim Mantle dead. However, it looks like from now on Mantle will become an API for select game developers that collaborate with AMD and want to use exclusive capabilities of Radeon hardware before they are supported by Vulkan, OpenGL or DirectX. In short, future versions of Mantle will become even more proprietary than the current iteration because they will not be available for all.

Discuss on our Facebook page, HERE.

KitGuru Says: Proprietary graphics APIs is not something that can survive in the long term unless the platform they serve prospers and its owner can afford rapid development of its application programming interface. 3dfx Glide and S3 Graphics Metal provided loads of advantages for appropriate graphics chips, but ATI and Nvidia, who relied on DirectX and OpenGL, managed drive both companies and their APIs out of business relatively quickly. Generally speaking, investing in Mantle is clearly not the best thing that AMD can do these days.

Become a Patron!

Check Also

Horizon multiplayer spin-off coming 2025, job listing suggests

The long-awaited Horizon Zero Dawn multiplayer spin-off could finally launch in 2025. Job listings suggest the studio is in the final stages of development.

49 comments

  1. AMD did best possible from worst possible conditions-
    Worst conditions-
    1. Many people still thinks why microsoft did not gave dx12 like multhreading capabilities to previous directx real ans is black nexus of microsoft and intel. Intel pays ms to tailor dx such that dx will rely heavily on single core performance no matter if u have 8 or 16 core amd cpu dx will always give winning chance to intel products
    2. MS tries to stop khronos group for making api which can challange dx coz most teens not use linux bcoz of no AAA title for opengl

    Mantle aftermath –
    1. MS feared badly that if they continue their unethical practice then eventually amd will realease mantle on linux and that will be hells gate for ms coz who will buy 200$ new windows if we can play all latest AAA TITLES on linux like on steam os than everyone will use windows7 for dailywork and linux for gaming , this will result in epic fail of upcoming windows…..thatswhy ms will give windows 10 for freee …….ha ha ha i told their closed door politics
    2. Open gl was pathetic when it comes to productivity and devs cried and died for newer api but khronos group and amd finally did check mate move coz mantle is now soul of vulkan api simply means whole world will use that, some famous opengl users like next playstation, nasa, whole automobile sector, animation and multimedia industry so now everywhere we will have a absolutly powerful api that can scale up fiercely to any level leaving dx in dust

    Devs now plz write AAA games for linux and i will live with win 7 no win 10 upgrade even if its free coz win 10 hv numerous loopholes for spying and selling personal information

  2. Søren Chr. Nielsen

    While it’s sad to see Mantle go this soon, it’s great that its existence led to the birth of Vulkan. AMD made the best possible decision here and I believe that by providing support to the API from the start, will give them an advantage untill Nvidia finally follows suit.

  3. Etienne Boutet boucher

    i agree with you except im not sure about your loopholes on win 10 except if you got good sources about it.

  4. AMD: FreeSync will be the next thing we put on the chopping block

  5. nVIDIA doesn’t have to follow anyone. Do you really think Vulkan will get more attention than DX12? If you do, you’re high on crack. As for it being AMD’s plan all along to dump Mantle for Vulkan, wouldn’t it make more sense for AMD didn’t say that when it came out. Why? didn’t they? Because they wanted you to go for Mantle because that’s all they had! AMD dumped Mantle because they knew they couldn’t support it alone. AMD doesn’t have money to burn, only customers.

    The fanboy backpedaling is real….

  6. Seriously; what is your point? AMD realised Vulkan was en-route to bringing the advances that Mantle was. They contributed to the project, and have effectively made parts of Mantle open-source. As for Nvidia, they’re already in the Vulkan boat. Scroll to the bottom on the Khronos website (https://www.khronos.org/vulkan) and you’ll see them among the “working group participants”, alongside AMD, Intel, Oculus VR, Sony, Apple, Unity, LucasFilm; the list goes on. You claim AMD fans are back-pedalling, but AMD has not made a wrong step. They’ve made a very right one, that will hopefully lead to better cross-platform standards.

  7. My point is DX12 will still be the main focus for BOTH camps. Look at OpenGL, if open standards were king, it would of took over for DX by now, but it didn’t. Vulkan will be NO different. Mantle wasn’t even all that good. The idea was good, but the execution sucked.

  8. “…when AMD introduced Mantle in 2013, it came as a major surprise. Nonetheless, it was a logical thing to do.”

    Well yea, when you stop competing with Intel on the CPU front, you find a way to shift consumers’ focus elsewhere. Especially when a 2.5GHz Intel CPU is equal to a 4.5GHz AMD CPU in games. If Mantle was never long term, AMD should of came right out and said, hey we’ve been working on this API called Mantle, but in the long run we want to partner with someone and do something better.

    Mantle sucked and AMD knew it. Enter Vulkan.

  9. Søren Chr. Nielsen

    So if someone doesn’t agree with your point of view, then they must be high on crack?

    I won’t bother making an actual counter-argument to what you said, since you didn’t bother to construct a proper argument to begin with.

  10. well DX12 is windows 10 only and vulkan is probably going to go back to windows 7 or earlier as well as have full support on linux based operating systems like steamOS

    vulkan is going to have the larger install base and with the aforementioned steamOS gaining a bit of steam (he he) more developers might end up using vulkan on steamOS and just simply use it again on the windows versions (because they are going to have to use it on windows 8 and 7 as DX11 will not perform in the same way)

  11. Mantle was BS, none games (maybe excluding BF4) benefited from it. They try to convince everybody, that Vulkan will be great thanks to their solutions – pathetic claims.

  12. i hope Vulcan can run in Linux and windows 7 and XP

  13. The Shiller (D W)

    “AMD needed an API which could exploit its GCN [graphics core next] architecture and would boost performance of Radeon hardware compared to Nvidia GeForce graphics cards in select games.” – This is misinformation. Mantle was developed to improve AMD CPU performance on games on Microsoft Windows.
    It cannot possibly improve fill rates, which is the main bottleneck on GPUs.

  14. Like I said…

  15. Ok.

  16. you said that anyone who thinks vulkan (an API that is to be on multiple operating systems including windows linux and probably OSX) is going to get more attention than DX12 (an API that is to be on windows 10 and windows 10 only) is high on crack

    i simply pointed out that vulkan will have the larger install base and the fact it will be on steamOS/linux should see it use much more than DX12 initially

    my reasoning for that is as follows

    there are many people still on windows 7 because of windows 8s start screen

    any developer using DX12 might have issues getting their game to work on DX11 for the windows 8 and 7 versions and not supplying DX11 versions would remove much of their potential audience

    so it would make business sense to use vulkan because it will be on windows 7 8 and 10 as well as for games that are put on linux linux

  17. Arafat Zahan Kuasha

    Everything except Windows uses OpenGL.. Android, IOS, Linux, Mac.. They’re all gonna use Vulkan the same way. Only interesting thing to see is how long Windows can hold up ..

  18. You know what GPU & CPU is in XboX 1 (DX11.1) ? OK maby you dont know:
    Its Radeon GCN and Jaguar CPU x8 (4Modules 2 cores each = 8cores)
    And do you really think nV is responsible for DX12 early Build? answer is no, 85% of the Code is from AMD Mantle ! so The DX12 is really a Mantle based API to boost Games in XboX 1 (important to MS Sales cuz PS4 has 75% more market now !) + PC.

    MS need that $ from XboX and DX12 (Mantle) is something they need badly 😉
    So the XBox1 can run Games in FHD min. 30FPS 😀

    Im not a fanboy, i have open eyes and thats all. Greets bro 🙂

  19. Vulkan is quite literally built upon the work that Johan Andersson et al. at DICE/AMD put in to create Mantle. That isn’t even a claim; it’s known fact within the development community.

    Mantle proved a specific point to not only IHVs, but to the likes of Microsoft and ARB/Khronos too. Vulkan is the natural progression on from Mantle and was always likely to succeed it if all went successfully, which it pretty much has done. Building upon Mantle’s open (soon to be publicly viewable) spec, and significant input from select engine developers in the industry (specifically Valve, Epic, Unity), we have ‘Vulkan’.

  20. Y U MAD? You should thank them for calling the future a bit sooner than Microsoft/Kronos intended. Most of the times it seems that AMD, even though they are in a financial pickle, push great standards to the industry.

  21. Sorry bro but can’t talk about loopholes in public forum

  22. Nope, it’s open sourced already. It’s part of DP 1.2a.

  23. Wrong. It’s AMD’s tech that ONLY works with AMD GPU’s.

    http://support.amd.com/en-us/search/faq/219

  24. As a former IT developerconsultant on Ms Windows systems the problem with DX12 is that it requires a massive piece of malware to run it. (since I retired ALL my computing devices are now Linux based or derived)

  25. lol good luck with Linux. You’re gonna need it.

  26. DX is doing fine. Seems the only time complaints come up is when open source fanboys see something that could POTENTIALLY be better when history hasn’t proven that to be true.

  27. DX12 is not Mantle. The “best parts” of Mantle went to Vulkan. Nice try tho.

  28. you said that anyone who thinks vulkan (an API that is to be on multiple operating systems including windows linux and probably OSX) is going to get more attention than DX12…

    I said the opposite. Learn to read.
    Backwards compatible or not, the difference being DX is what 90% of are using. Btw, new computers will come with DX12 out of the box. So you think devs are going to abandoned DX for an Open Source API? Look at Mantle vs DX. DX was threatened and M$ upped their game (with the help of AMD and nVIDIA mind you) with DX12. You can’t stop the big guys with Open Source shite. Look where open source got Linux on the desktop versus Windows.

  29. AMD doesn’t push anything except clock speeds on rebranded GPU’s.

    See: 200 and 300 series Radeons.

  30. Dumped Windows for Linux over 2 years ago – never regretted it for a moment – networking issues connected with tethering were a nightmare on Windows eg I would get a signal drop and when restored Windows would simply fail to reconnect or take ages to do so – in Linux reconnection is automatic and almost instant. I also don’t have to worry about the latest sneaky attack getting through layers of anti-malware eating up cpu cycles and slowing my system down. The only downside is that I cannot (yet) play Fallout Las Vegas on Linux but on the plus side SteamValve has now established a Linux prescence and there are a growing list of attractive titles that play right out of the box.

  31. Oh Len…

  32. Even if new computers come with Windows 10, that won’t change the fact that there will be users of windows 7/8/8.1 NOW that wouldn’t want to suddenly change now, and move to 10.

    While i understand that Windows has a large install base, when it comes to DX12 adoption, that doesn’t matter much since it’ll only be available only on windows 10. That divides the user base between Windows 7/8/8.1, and Windows 10. Meanwhile, Vulkan is not only supported on Windows 7/8/8.1 (and even 10), it is also supported on linux, mac, and android. If a developer has an option to, why would they make a directx 11 AND 12 version of a game (respectively for windows 7/8/8.1 users, and windows 10), when they could make it on Vulkan, which already supports all 3 versions of windows?. When you think about it, wouldn’t that give Vulkan the advantage for adoption? It would.

    Vulkan as an api is designed more like Direct X, giving developers a simple api that should be within their comfort zone. that means that unlike the OpenGLvs Direct X situation in the past, Developers wouldn’t have any real reason to pick Direct X over Vulkan. Combined with universal OS support (including windows 10), it would seem as though Vulkan should be the better choice.

  33. So when DX10 was introduced with Vista, devs ditched DX for OpenGL? Oh right…

  34. Vulkan is targeted for mobile too.
    Nividia will want mobile too.
    Result is clear.

    And besides Vulkan may be based on Mantle, it was created by few dozens of companies. Nvidia included.

  35. but you seem to forget the outdated opengl drivers placed within vista, and also the other misinformation about opengl (http://blog.wolfire.com/2010/01/Why-you-should-use-OpenGL-and-not-DirectX ). Even so, that can’t really apply to Direct X vs Vulkan, because like i mentioned before, Vulkan is designed like Direct X, so there is less reason to stay with Direct X.

    And when Direct X 10 was introduced for vista, they were already using DX9. For DX10, Vista wasn’t worth it, so they continued to use DX9 (which had more support across windows OS, due to multiple Windows versions). And that, again, goes back to how DX was easier to implement than Open GL, which again, doesn’t apply to Vulkan, cause they have an option without that downside.

  36. Oh god, you’re insane.

  37. Um nVIDIA doesn’t want mobile. Maybe you should look on the net for the plans nVIDIA has with their X1 offering… I’m almost getting tired of schooling you scrubs.

  38. I don’t know how making that assertion serves to make a point, especially with the topic but, do what you will.

  39. spending time with all this schooling, he barely has any time for selfies anymore. sheesh get a clue people.

  40. nvidia mobile chips… shield…. yep, doesn’t want mobile.

  41. That’s their gaming and home entertainment market. nVIDIA is bailing from phones and tablets, and focusing on games (Shield), Android TV, and automobiles.

    FYI, Just because something is the size of a phone or tablet doesn’t put it in the mobile category…

  42. Arafat Zahan Kuasha

    As long as Windows does fine, DX will do fine

  43. Bingo.

  44. amd has been working with microsoft to build DX12 using what they learned from the consoles and mantle update what you know cause it’s all wrong

  45. You’re aware that Nvidia is one of the Khronos partners on Vulkan right?
    Lol, who’s a fanboy now?

  46. vulkan copy and pasted the best bits from the mantle blueprint, nvidia is already on board so is Intel and a alot of game studios from the likes of epic, blizzard, valve…

    you can go and compared vulkan (vk) to mantle (gr) below

    vkBeginCommandBuffer == grBeginCommandBuffer
    vkCmdDbgMarkerBegin == grCmdDbgMarkerBegin
    vkCmdPipelineBarrier == ???
    vkCmdDbgMarkerEnd == grCmdDbgMarkerEnd
    vkEndCommandBuffer == grEndCommandBuffer
    vkCreateFence == grCreateFence
    vkCreateBuffer == ???
    vkGetObjectInfo == grGetObjectInfo
    vkQueueSubmit == grQueueSubmit
    vkAllocMemory == grAllocMemory
    vkBindObjectMemory == grBindObjectMemory

  47. You haven’t noticed certain Mantle enabled game engines was quickly converted into Direct12X version.
    A hint for you
    https://pbs.twimg.com/media/CBBu9COWwAAPzZB.jpg

  48. That’s referring to a specific implementation which is based from the available standards.
    DP1.2a is part of DP standard. DP1.2a also part of DP1.3.

  49. Just like your bs. is unreal…
    If AMD didn’t try to go for mantle we would be pushing DX11 forever.
    How did AMD burn us??? looked at some DX12 benches lately.
    Try finding out how DX12 runs on an FX 8350 compared to DX11.
    DX12 learned from mantle, mantle leads to Vulkan. That’s 2 wins for me running a complete AMD system. Thank you AMD for pushing the industry