AMD is spending more resources making slides about Mantle than actually getting it working in games. >_>
Heh. I hope that you're kidding. If the game publisher is smart, they are waiting to see if Mantle's performance benefits are actually worth the development effort before advertising the hell out of it.
I only wish that AMD's marketing department took the same tact. I personally don't want to see another tweet, press release, or slide deck from them related to Mantle until they actually have the drivers ready for end users to download. I think that most people would agree with this.
You know what? I'll go one better... I think that the Mods should lock this thread until Mantle is released. It would certainly cut down on the trolling and rampant speculation.
+1 - So tired of the slides etc etc....its time to show us what you got AMD and co.
While I share that frustration, and wish Dice would release this thing, "for crying out loud," I don't see any point in locking this thread. Maybe rename is to "the Mantle speculation thread." But I enjoy the speculation and anticipation, and hope it succeeds. But if it falls flat, oh well.I think that the Mods should lock this thread until Mantle is released. It would certainly cut down on the trolling and rampant speculation.
What I don't understand is those who want to rain on the parade, What's that about? How is that entertaining? And if Mantle does fail, as many of you seem to hope, how do you benefit? I haven't figured that out yet.
While I share that frustration, and wish Dice would release this thing, "for crying out load," I don't see any point in locking this thread. Maybe rename is to "the Mantle speculation thread." But I enjoy the speculation and anticipation, and hope it succeeds. But if it falls flat, oh well.
What I don't understand is those who want to rain on the parade, What's that about? How is that entertaining? And if Mantle does fail, as many of you seem to hope, how do you benefit? I haven't figured that out yet.
Must mean they canceled mantle. Mantle cancellation confirmed.
The fact remains that if Mantle succeeds, it will require 2 builds for every game that includes it. A Mantle build and a DX build for the other 95% of the people who will play the game.
AMD is spending more resources making slides about Mantle than actually getting it working in games. >_>
http://www.joystiq.com/2014/01/20/thief-pc-specs-slip-out/?ncid=rss_truncated
http://store.steampowered.com/app/239160/
Thief specs listed. Mantle is funny enough missing. Both DX10 and 11 supported.
No. First, I believe you mean releases, not builds.
Second, It will require an additional code path for the Mantle calls instead of DX calls, and additional engine code for the Mantle stuff.
It certainly is not going to require a completely different release of the game just because of a different renderer code path.
It would be much easier from a development standpoint to have two alternate builds rather than all of that packed into one.
IIRC, Johan said the Mantle version of BF4 would likely require a completely different build, which could be expected. Having so many code paths completely different (and the logic based around what each is doing) makes little sense to bloat each release. It would be much easier from a development standpoint to have two alternate builds rather than all of that packed into one. The installer can easily have the option between which version to install.
Release in terms of a separate product, packaging and such? No. A build based on source code from a different branch being included? I can certainly see it requiring such.
And regardless of that, you still have to fully test the game using both rendering paths (even if it is all in one build). That requires double the QA tests.
You've never done large scale software development. Doing it your way would be insane. I can't think of a single reason to do it that way. I can't even think of any SDLC tools that would operate in such a fashion.
You've never done large scale software development. Doing it your way would be insane. I can't think of a single reason to do it that way. I can't even think of any SDLC tools that would operate in such a fashion.
I have done large scale software development. It isn't hard to bundle two executables into the installer and have each choose which will be installed.
And I can't think of a single SDLC tool that is popular that doesn't allow for multiple branches of files.
What is most likely happening in the BF4 development is they have a trunk of DX code and a branch that contains all the Mantle. Once they are satisfied with both, they can merge them together and include some API to determine which path to use (DX or Mantle), adding in some performance hit, or they can simply build using the DX source and then build against using the Mantle branch source. Include in the installer both executables and the assets required to run the game.
And I can't think of a single SDLC tool that is popular that doesn't allow for multiple branches of files.
I have a few games which include multiple .EXE's for different DX versions. The ones I have, or know about, include both .EXE's during install, and let the launcher call the one you choose with your settings.
That would certainly be a logical way to do it.I was thinking Mantle would have different assets at first for some reason, but in reality, it would use 100% of the same as the DX version. So, simply including another (Mantle specific) executable would be trivial with two builds.
A branch isn't a whole new release. That's why you would have one project, and the different rendering paths are different packages. Then the rendering paths come together when your code is promoted to merge state.
You certainly don't maintain two entirely separate code bases for a game.
IMO mantle is a mistep. AMD should have spent the time developing a physics engine for use on their igps, making it easier on the CPU.
Video has not been made available in my country.
I said two different builds.