AMD has revealed an API that gives developers direct access to GPUs using ...

General Discussions
#21

I used to be extremely skeptical to Mantle since I thought Nvidia wouldn’t be able to implement it. If Mantle failed it’d be a loss for AMD which could increase Nvidia’s market share killing competition, and if it succeeded it’d split up PC between Nvidia and AMD since in that case Nvidia and Intel might make their own Mantle-like API. If Nvidia actually swallows their stupid pride and makes a Mantle driver (assuming they can) I would definitely start using it immediately. Multithreaded rendering? Precompiled pipelines? Much more flexible state? Just to mention a few? Sign me the f*ck up.

Does anyone have a clue on how hard it’d be to create a Java binding?

#22

Without having seen any code, I’m sure I’ll add it to LWJGL the same day it’s out. Interestingly, we might have to start worrying about JNI. It’s only noise with OpenGL, but if we’re moving to 100k-1M draw calls per frame, JNI overhead could easily add up to a few milliseconds. An FFI mechanism would be nice in Java 9…

I think a multi-vendor, multi-platform Mantle would be a great opportunity for Linux to directly compete with Windows as a gaming platform. I also can’t wait for lightweight drivers vs the 200+ MB monstrosities we have to download now every couple of weeks.

#23

Damn. Have a medal, good sir.

#24

Biggest hurdle for Mantle is going to be getting Nvidia, Intel and Apple to support it, without that not sure its going to have widespread enough appeal. For that to happen it’ll need to be an open standard and AMD will have to give up its exclusively control on the API probably to a party like Khronos.

Having said that, AMD is in a pretty strong position atm as their chips are used in all the next gen consoles (PS4, XB1 & WiiU) and in about 1/3 of the gaming PC’s out there. Therefore it could become pretty common as the code is suppose to be shareable across hardware.

#25

Read this article: Tech Giants […] Join the HSA Foundation

HSA seems to be an architecture (or concept?), that allows CPU and GPU to access the same adress space, i.e. passing pointers instead of copying data, and other stuff. The new AMD third generation APU (Kaveri) A10-7850K will have this. Kaveri will include AMD’s Mantle API [2].

The article above says that many companies, including Oracle, are members of the HSA Foundation. It says also, by joining HSA Foundation Oracle strives for better utilization of the GPU with Java.

Other interesting things:
[1] Java 8 Sumatra processing demo (Sumatra is an OpenJDK Project to allow to take advantage of GPUs and APUs)
[2] a biased article with more details about the APUs (read also comments of user Ken Luskin for more background info)

edit: Oh, I see there is already a thread about HSA+Java.

#26

Nvidia and java have announced they will be allowing speed boosts using the gpu , this involves even array scanning functions with up to 48* performance on some tasks.

#27

Missed that but seems like its acceleration is done using Nvidia’s CUDA library.

Its not clear from the links I’ve found whether it’ll be automatic behind the scene acceleration, like what AMD is working on for Java 9 (using OpenCL), or whether its just a simple binding to Nvidia’s CUDA libraries which developers will have to work with directly, since the articles states Java 8 then i’m assuming its probably just a binding to CUDA.

Link 1 and Link 2.

#28

Noticed the following new video presentation in relation to Mantle:http://www.youtube.com/v/QIWyf8Hyjbg

source

#29

very interesting vid.

now will we be able to use this while leveraging java… :slight_smile: oh i hope so…

#30

Merry Christmas and a happy new year! Please accept this present from me and my fellow Swedes!

#31

This is an excellent point, and I really wish that’s what would happen, but here are the best possibilities I see:

  1. The only titles that take advantage of this technology will be console-centric games where the publisher has decided that it wants to have the option to release quick and “less dirty” ports to the PC. This publisher will have decided that it wants to be the first to make its games inaccessible to much more than half of the primary PC gaming market–or roughly half, depending on how significant Intel is in your considerations (but growing day by day). In order for the publisher to be motivated enough to do this, AMD would need to somehow incentivize the publisher’s adoption of this business strategy. The PC is not a closed platform like a console, however, so fewer options are available. AMD would basically need to pay directly out of pocket for a blockbuster game–something it will not do because gaming does not make up enough of its bottom line.
  2. A title using this API will be in need of the performance enhancements it offers… because the current generation consoles have hit their limits already…? The only case in which this would be relevant would be when a publisher was releasing a title to all four platforms (Xbox, PS, Wii, restricted PC), wanted its game to look as cutting edge as possible, and needed the Wii U version to be visually consistent with the others… without sacrificing all of the advantages of running on current gen hardware in the process.
  3. I forgot what three was, but you can be sure that it was as important (and unlikely) as my previous two points.

I am a big AMD supporter, but Intel and NVidia–really just Intel–don’t have to do anything at all, and they know it. How does all of this affect Java games development? The only way this could possibly affect it is if Java code ran on any of the games consoles, which it doesn’t the last time I checked.

#32

@Ghida Mantle is a PC exclusive API atm (and probably SteamBox) and its not going to be available on any of the current consoles (PS4, XBO, WiiU). Sony & Microsoft already have their own low level api’s available on their consoles which give features similar to what Mantle now provides PC.

#33

First real-world results: Mantle renderer now available in Battlefield 4. Up to 58% faster in a high-end configuration.

#34

Nice improvements.

Interestingly John Carmack mentioned awhile back that he reckons Nvidia’s OpenGL extensions (not sure which ones in particular) can give similar improvements. If so and such changes are incorporated into future OpenGL core then could be a problem for Mantle adoption.

#35

When can we see LWJGL support? I have my Mantle dedicated computer ready to be put together. xd

#36

I’m sure new contributors are always welcome. :wink:

#37

You must be referring to this (this is all cross-platform btw, not Nvidia specific). Unfortunately, like Mantle, it’s very limited in terms of hardware and driver support. And also requires a major rewrite of the rendering pipeline. I think it will all come down to availability and driver stability. Technically, it should be easier to write a robust driver for something like Mantle vs the beast that OpenGL is.

[quote=“theagentd,post:35,topic:44368”]
Whenever AMD releases a public specification, which I guess won’t be soon. With the driver released today, Mantle will only work on R9/R7/Kavery parts. They’ll probably wait for GCN 1.0 support first and also for any tweaks that will come up from the BF “beta test”.

edit: The above is wrong. The driver will support Mantle on all GCN cards, but BF+Mantle will only run on GCN 1.1 (for now). Still no info on the Mantle spec.

#38

Mantle functions, the list.

#39

and an attempt to apply Mantle’s design to OpenGL.

#40

Lol. gl* —> gr* The Japanese won’t be able to differentiate between them when talking… xD Also, GRRRRR! Angry functions!

Cool list! Seems pretty simple actually!

I’m not sure I’m getting what he’s trying to do? Redesign OpenGL to look more like Mantle? Would that actually bring any of the (performance) advantages of Mantle over to OpenGL?