This is true, DX10 started a trend to gradually replace the legacy fixed function style pipeline to a better fit for modern GPUs.
I have seen that over the years generalization tends to come in waves so I doubt this will be the last upset we see
This is true, DX10 started a trend to gradually replace the legacy fixed function style pipeline to a better fit for modern GPUs.
I have seen that over the years generalization tends to come in waves so I doubt this will be the last upset we see
Both latest DirectX and OpenGL versions are updated to reflect new changes in hardware. It's not like they are stagnating, and it's not like any new standard will be adapted in one moment.
Peteris.
Interesting that Star Citizen will fully support Mantle.
http://www.guru3d.com/news_story/amd_mantle_graphics_api_adopted_by_various_developers.html
I'm due a graphics card upgrade in a few months; Very interested to hear if E : D in alpha runs better for nvidea or AMD...
http://www.guru3d.com/news_story/amd_mantle_graphics_api_adopted_by_various_developers.html
I'm due a graphics card upgrade in a few months; Very interested to hear if E : D in alpha runs better for nvidea or AMD...
Hmmm, the plot thickens. This isn't making my next choice of GPU any easier you know
I just bought Nvidia for the first time in years... It would be churlish of me to wish that Mantle doesn't provide a big performance gain, wouldn't it?
Nvidia do have G-Sync on their horizon, which sounds very interesting.
To the right.
floor(dx + 0.5).
Philip Coutts
Volunteer Moderator
No doubt Nvidia will come up with something similar (G-Sync?) even if they don't the game will run just fine on their cards.
+1, Just burst out laughing in the office
Yeah I like the look of G-sync, but it is a pain that you have to buy a new monitor for it. Still continuous improvement is what PC is all about.
It makes me sad to see this kind of backwards movement in the industry
Someone else already mentioned that DX and GL are not in sync with the level of abstraction that we use in games today, and that, above all other things, should be what a new API addresses if there were to be one. PC devs are in rough, rough shape already - one API is literally a state-machine C API, the other supports only 1 out of the 3 major OSs. Sad times.
And now...introduce a lower-level API? Come on. Development time is the major bottleneck in modern games, not the draw calls. You do not move to a more hardware-specific architecture to help the already-absurd video game budget problem. I could get on board with a new API if it advertised "cross-platform, strong abstraction, clean & natural C++ API." That's something I would get excited about. But "low-level" and "close to the metal"? Not the words I wanted to hear...
Really seems a whole lot of graphics programmers out there have their heads in strange places (this coming from a graphics programmer....)
Anyway...hope FD does the right thing and steers clear of this
Someone else already mentioned that DX and GL are not in sync with the level of abstraction that we use in games today, and that, above all other things, should be what a new API addresses if there were to be one. PC devs are in rough, rough shape already - one API is literally a state-machine C API, the other supports only 1 out of the 3 major OSs. Sad times.
And now...introduce a lower-level API?
Come on. Development time is the major bottleneck in modern games, not the draw calls. You do not move to a more hardware-specific architecture to help the already-absurd video game budget problem. I could get on board with a new API if it advertised "cross-platform, strong abstraction, clean & natural C++ API." That's something I would get excited about. But "low-level" and "close to the metal"? Not the words I wanted to hear...Really seems a whole lot of graphics programmers out there have their heads in strange places
(this coming from a graphics programmer....)Anyway...hope FD does the right thing and steers clear of this
Interesting half (or "quarter"
) step in the right direction (I think), but afaik is purely HW dependent only on monitors. A sort of lightboost v3. Hope it turns into an industry standard, or a similar version.Anyway I'm surely not going to trade down my 27" IPS Dell monitor for some Asus TFT thingy anytime soon, but it could be a factor on future monitor purchases. A G-sync 4k, passive 3d enabled, 120hz 3d monitor would be interesting
Last edited:
It makes me sad to see this kind of backwards movement in the industry
Someone else already mentioned that DX and GL are not in sync with the level of abstraction that we use in games today, and that, above all other things, should be what a new API addresses if there were to be one. PC devs are in rough, rough shape already - one API is literally a state-machine C API, the other supports only 1 out of the 3 major OSs. Sad times.
And now...introduce a lower-level API?
Really seems a whole lot of graphics programmers out there have their heads in strange places
Anyway...hope FD does the right thing and steers clear of this
The logic and viability behind Mantle depends on the consoles (and their more similar than ever CPU and graphic subsystem). If you develop for consoles you will be working at a lower level (at Least some point along the console lifecycle). For what is known of Mantle is that it proposes to devs to keep more or less the same already optimized code for the PC port, as long as the system has a GCN based graphic architecture.
It does not preclude the need for a general direct X version.
Last edited:
Hmmm, the plot thickens. This isn't making my next choice of GPU any easier you know
For me it is easy. Nvidia. Because they have better support for BOINC. I crunch data for SETI@home (but also Climateprediction and Milkyway@home) and am in the top 5% of the world. With AMD I had to install all sorts of junk to make it work.
I don't know what all this Mantle will do but right now there is not a lot of difference to my eyes between both brands.
I know Chris Roberts does not like DX.
Yeah, I'm confused by this. On the one hand we're talking about "low-level," "on the metal," etc. On the other hand we're talking about portable console code. But....which does Mantle want to be? Is it high-level or low-level?
If it's an attempt to unify console graphics architecture with PC, bring all GPU code under a common, portable API, then I can get on board with that. But why is it not advertised as such? The focus of the advertisements thus far is "get the most out of your PC" "thin drivers" etc. Everything I read is talking about performance - the performance gains of eliminating an abstraction layer and talking directly to the card, the performance penalty of "high-level APIs like dx & gl" (ha), etc.
Well, bottom line, I'm certainly not against it for the sake of being against something...if it decreases development cost, I'm in, but that doesn't seem to be what it's trying to do. I guess we'll have to wait for more information to have a serious discussion though, as virtually nothing except hype has been revealed thus far. But if what they're doing here is actually creating a beautiful new cross-platform graphics architecture, they sure as heck aren't using the right words to describe it
Looking forward to more technical details
You're looking at it from a developer/developing perspective... try looking at it from a consumer perspective - more better graphics for no more monies! Then look at it from a developer/marketing perspective...
Yeah, I'm confused by this. On the one hand we're talking about "low-level," "on the metal," etc. On the other hand we're talking about portable console code. But....which does Mantle want to be? Is it high-level or low-level?
If it's an attempt to unify console graphics architecture with PC, bring all GPU code under a common, portable API, then I can get on board with that. But why is it not advertised as such? The focus of the advertisements thus far is "get the most out of your PC" "thin drivers" etc. Everything I read is talking about performance - the performance gains of eliminating an abstraction layer and talking directly to the card, the performance penalty of "high-level APIs like dx & gl" (ha), etc.
Well, bottom line, I'm certainly not against it for the sake of being against something...if it decreases development cost, I'm in, but that doesn't seem to be what it's trying to do. I guess we'll have to wait for more information to have a serious discussion though, as virtually nothing except hype has been revealed thus far. But if what they're doing here is actually creating a beautiful new cross-platform graphics architecture, they sure as heck aren't using the right words to describe it
Looking forward to more technical details
Well, you are absolutely right in saying that data about Mantle is scarce. I wouldn't call hype (but for the usual suspects) but more of speculation going around. And some of what I post is also speculative, although not really in the hype type (I think).
We know that Mantle a GCN based graphic architecture. Having as target a HW family and not a precise model (e.g. consoles), it will not be as low level as some other implementations of past, but compatible accross all GCN based HW. So far, so good and not very relevant... if GCN was not only the basis of all modern AMD cards (since 2 years ago) but all the real "jackpot": PS4 and XBone.
The low level languages (or CTM) of those machines will not be called Mantle: Direct x in MS console (I think) and some other name in PS4. But the question is... how different are those CTM languages from Mantle? Targeting the same shader & all type of HW, the differences could be pretty slim, and optimization rules the same.
Now without Mantle, the dev would have to convert that graphic code into DX or Open GL. Wasting quite some time and resources and, quite significant, losing optimizations. If a very similar graphic environment is available on the PC, the dev wouldn't have to do that extra work and would have a better end result. And that environment is Mantle.
Some questions arise though - not all GCN devices are equal (290x vs HD7750), and will become less equal as GCN evolves. Well "under" Mantle there is a thin driver to ensure compatibility. Like there are (quite fatter, I presume) drivers for Direct X and OGL.
And why would devs be interested in "playing the Mantle card?" and not just keep business as usual?
Well, for two things: if Mantle allows them a degree of control and performance that no other "language" allows and if it actually helps to automate the porting between platforms.
According to the testimony of the Dice guy Mantle can have a significant impact and was build with the developers and their requirements in mind. It has HLSL "hooks".
And so far, as said, details are scarce, and the devil lies in the details. So it is a bit of "wait and see" before any conclusion can be drawn.
That is the way I see it - although my actual knowledge of graphic pipelines and engines is far far far too lacking to make any solid analysis of viability/evolution.
Regarding the main question if ED should/would support Mantle, it depends on how easy it would be to incorporate it in Cobra engine and what impact it would have on the game graphical performance. As we know neither of this elements, we will have to wait for FD to say something, although I think they also can't say anything about either question so far.
It makes me sad to see this kind of backwards movement in the industry
Someone else already mentioned that DX and GL are not in sync with the level of abstraction that we use in games today, and that, above all other things, should be what a new API addresses if there were to be one. PC devs are in rough, rough shape already - one API is literally a state-machine C API, the other supports only 1 out of the 3 major OSs. Sad times.
And now...introduce a lower-level API?
Really seems a whole lot of graphics programmers out there have their heads in strange places
Anyway...hope FD does the right thing and steers clear of this
I totally and utterly disagree with you. Direct3D is not wringing out sufficient levels of performance out of our GPUs. You only have to look at the PS4/XB1 GPUs to realise that. There needs to be a lean and mean API. Why should consumers spend their hard earned money to have OS system software screw it up. Look at Sony and what they extract with libGCM. MS aren't really that bothered with PC gaming since the Xbox. SteamOS I will be migrating to as they seem to be more interested in PC gaming and a protest vote. MS I'm gonna give you a bloody nose
Windows 8, Surface and phone MS you're gonna have a big bloody nose. Maybe nVidia and AMD should collaborate on a separate high-level API. If this lower level (note I say lower) API can deliver more performance on same hardware why not. From the looks of it is going to be embedded in quite a few engines.
Yeah, I'm confused by this. On the one hand we're talking about "low-level," "on the metal," etc. On the other hand we're talking about portable console code. But....which does Mantle want to be? Is it high-level or low-level?
If it's an attempt to unify console graphics architecture with PC, bring all GPU code under a common, portable API, then I can get on board with that. But why is it not advertised as such? The focus of the advertisements thus far is "get the most out of your PC" "thin drivers" etc. Everything I read is talking about performance - the performance gains of eliminating an abstraction layer and talking directly to the card, the performance penalty of "high-level APIs like dx & gl" (ha), etc.
Well, bottom line, I'm certainly not against it for the sake of being against something...if it decreases development cost, I'm in, but that doesn't seem to be what it's trying to do. I guess we'll have to wait for more information to have a serious discussion though, as virtually nothing except hype has been revealed thus far. But if what they're doing here is actually creating a beautiful new cross-platform graphics architecture, they sure as heck aren't using the right words to describe it
Looking forward to more technical details
It's lower level than Direct3D but not low-level like assembler...