“Khronos Group, the association behind OpenGL, has today announced the fourth generation of its cross-platform API spec, which takes up the mantle of offering a viable competitor to Microsoft’s DirectX 11. The latest release includes two new shader stages for offloading geometry tessellation from the CPU to the GPU, as well as tighter integration with OpenCL to allow the graphics card to take up yet more duties off the typically overworked processor.”
4.0 is the first real step up that is going to lead to much better competition with DirectX. This is the “is ready for the desktop” (as the Linux fans say) release. Let’s keep up the momentum with OpenGL (and it’s technologies). Better, Faster, Stronger. w00t
How exactly where OpenGL 1.x, 2.x, and 3.x not “ready for the desktop?”
the “ready for the desktop” term was an example of a saying specifically aimed at linux users.
OpenGL has had many releases that, while good, still didn’t compare to DirectX as far as features and capabilities relating to gaming. But ever time a new release came out everyone got excited as if it would finally be in a position to be as good or better than DirectX.
NOW! it looks like that time is really about to happen. I am not going to get into all the reasons why, but if you look into everything that has gone into this 4.0 release you will see.
So to sum it up for you; I used (in quotes no doubt) “ready for the desktop” as a play on words to draw a parallel to a similar situation. really ready in an equal to or greater than style fashion, to its rival DirectX. This is opposed to the linux saying i was quoting being equal to or greater than Windows (it’s common comparison.)
This is the first really big step to that, and the version 4 releases are going to be an amazing thing for the platform.
Edited 2010-03-11 23:48 UTC
Well, if you have to give such lengthy explanation it means that your joke wasn’t that good to begin with.
That being said, you are comparing apples to oranges. DirectX is a full multimedia API, while OpenGL is a 3D API. Technically it should be Direct3D vs OpengGL. And honestly, they have been both in real parity feature-wise.
OpenGL 4 is parity with the Direct3D component of DirectX 11. Which was just released, and is far from being the norm for windows desktops. So it seems both systems are pretty much neck to neck.
DirectX can be compared to OpenGL+AL+CL, but in that case, DirectX still Win. -But- the OpenGL+CL target is more oriented toward CAD niche / Render than modern gaming. It is not a bad thing, DirectX != Open**, they have a targeted market.
That said, more game on Linux would help the whole “Linux on the desktop” thing, dead or not. Most young gamer geek would switch, just to be different, it would push Linux in a new direction as a massive (>1%) market segment would at a good pace make the switch. They are young, they don’t care testing “new stuffs”.
OpenGL (!ES) is more dead than Bela Lugosi.
So were 2.0 and 2.1, as I recall. Of course, OpenGL only really targets Direct3D, not the whole of DirectX.
OpenGL is now doing a much better job of providing innovative, labor-saving features, that will allow it to be as good as, or maybe even better than DirectX, instead of merely “good enough to be usable, when you have to use it.” That’s the difference in 2.0/2.1 and 4.0 that PoundSmack was getting at. 2.1 was “good enough,” 4.0 is “appealingly and competitively great.” Hopefully.
Hopefully.
Isn’t it always the case though, that each release is great and all that?
I’m hoping that with more game developers focusing on mobile phones with OpenGL ES that they’ll be more willing to look at OpenGL on the desktop.
If someone were willing to develop an input library that would work across operating systems the way OpenGL works for output, it might loosen DirectX’s grip on developers.
When I first started using Linux the only reason i wasn’t prepared to use it full time was because of direct-x and games.
After I bought an xbox (ironic in itself i still bought into direct-x) I switched completely.
Microsoft doesn’t provide anything in direct-x that you shouldn’t be able to get in OpenGl except a vehicle to lock people in who still play games and it’s been working quite well for them.
Maybe I should have bought a Playstation instead but that’s another story. (I bought a 360 too.)
If that was the case, Windows game developers obviously would have used OpenGL instead. They are not all stupid.
Why in the hell did you link to Engadget?
http://www.khronos.org/news/press/releases/khronos-unleashes-cuttin…
The facts and rightful copyright is above.
Classy. Real classy.
Maybe because that’s the news article that was submitted to him? I kinda thought that the whole point of the “news” column was that it was just going to be an undigested reposting of news items that other people sent in.
If you want to be angry at someone, yell at pundsmack for not sending in the right link. Which is equally absurd, but at least properly directed.
I see this all the time, comparing openGL to DirectX, when it’s a completely unfair comparison for OpenGL since on it’s own, all it does is 3d rendering and texture-mapping, and that’s IT.
OpenGL is not 1:1 to directX, it’s 1:1 to Direct3d, a small subsection of the directX specification.
DirectCompute (What OpenCL is for) is just the tip of the iceberg… What about DirectInput which lets you enumerate and access input devices? DirectSound for audio playback? DirectDraw specifically for 2d work? DXGI for multiple display capabilities? DirectWrite for font handling? Shall I go on?
DirectX is not the weapon of choice of everyone except ID for no reason. It’s a full package of a consistent API all designed from the start to work together. Sure, running OpenGL on top of SDL will deliver a lot of that, but the API’s are radically different so you have to learn two ways of doing the same thing… and let’s face it, SDL’s audio is so anemic people will dive for OpenAL almost instantly.
Admittedly, OpenAL is so good even DirectX programmers will often choose it instead of DirectSound.
But you get the idea. DirectX is a powerhouse monster of multiple API’s that go so far beyond the scope of what OpenGL delivers it’s not a fair comparison. The only 1:1 comparison is to compare OpenGL to Direct3d – which frankly on the implementation side from hardware vendors you are lucky if you can make a openGL 2.0 program that’s real world deployable. 3 isn’t even really deployable yet and they’re talking 4?
Shades of the number two browser maker implementing CSS3/HTML5 when they don’t even have CSS2/HTML4 working right yet. (yes mozilla, I’m looking at you) – same thing.
In a lot of ways though that’s open source for you – constantly in catch-up mode with little to no hardware vendor support apart from lip service… at least until ID’s next big release which requires it – and then it will fade into obscurity again as three or four years go by without an ID release.
Hell, apart from ID the only reason we have OpenGL games on PC hardware at all is UDK – since it can target BOTH platforms…
Edited 2010-03-12 08:33 UTC
“What about DirectInput which lets you enumerate and access input devices? DirectSound for audio playback? DirectDraw specifically for 2d work?”
None of this is relevant for most modern games, except perhaps the input device stuff but this is hardly a huge and complex thing to do with other APIs such as SDL anyway (and the abysmally poor quality of input handling in most windows games doesn’t build a good case for direct input either).
Most games don’t use directsound directly but use a middleware such as miles sound system or fmod. Besides, most game developers don’t care much about audio.
Nobody uses directdraw anymore for the hud and the UI, people use direct3d to draw everything (and again, it’s something you’d get with SDL anyway on another platform).
Like it or not, direct3D is the only part of directx that game developers really care about.
Edited 2010-03-12 10:26 UTC
I’ll pretty much second that. And it’s worth noting, SDL is actually — in my never-humble opinion — a pretty good library on it’s own. I haven’t used D3D, I’ll admit, but I’m pretty happy with the tools I have available in SDL+OGL, and I like what I’m seeing in OGL4.
Maybe you should. DirectInput has been deprecated (replaced by XInput), just like DirectSound (replaced by XACT), DirectDraw, DirectPlay and DirectMusic (replaced by nothing). Direct2D isn’t much more than a simplified Direct3D.
Graphics and GPUs have been the sole focus since DirectX 9.0. DirectWrite, Direct2D, DirectCompute… It’s still about graphics.
The consistency of distinct modules doesn’t matter that much. Unless you’re dealing with an indie project, it’s quite unlikely that the audio guy will be dealing with graphics or network. Maybe it does becomes an issue if you didn’t really spend time to design your application…
directmusic has been replaced by Xaudio2
the whole directx thing is mostly being replaced by XNA & friends tho
Something that you overlook is that it is in response too, not in spite of user’s demands that the OpenGL standard is undergoing such rapid evolution. One of the major complaints about OpenGL was that it was not evolving quickly enough to keep pace with the changing demands of game developers, and the changing capabilities of modern graphics cards. People actually wanted to see OpenGL start growing more quickly — much like D3D, which has started to undergo change more quickly itself.
Part of the intent of that is also to simplify driver implementation, and improve driver quality. OpenGL drivers are of notoriously poor quality: one of the often-pointed-too reasons is that OpenGL is a complex API, with multiple, overlapping ways to do anything. This is because they’ve never deprecated anything, so up untill 3.0, lots of old, fixed-functionality OpenGL 1.0 calls where still available, cluttering up the API and adding complexity to the driver. So, again, deprecating those calls, and removing them from the core API, is an effort to create a cleaner, more orthogonal API and make everyone’s life that much easier. And that’s a good thing.
Edit: Oh, also, the OpenGL specification is not an open-source project, so it simply doesn’t reflect on the Open Source community. Unless “open source” really means “everything not Microsoft.” Then, sure.
Edited 2010-03-12 18:31 UTC
Directx is series of APIs that work in a perfect symbiosis together making it ridiculously easy life for the coders.
if Linux had viable choice for Directx as whole, it could easily chomp 50% of Windows’s marketshare within months of time.
Don’t forget “and viable video drivers that were easy to configure and delivered 100% of the hardware performance, instead of 60% and having to turn off all the ‘extras’ if you want to run more than one display.
But of course without a stable hardware API making hardware vendors WILLING to write drivers for it, instead of relying on people who just “feel like working on it out of the goodness of their hearts” that’s never going to happen.
Let’s be honest – Look at Windows 98 and MacOs 5, and they BOTH make X11 implementations look like a total joke when it comes to running more than one display at a time. You put 3d in the mix, and DXGI blows anything even PLANNED for OpenGL on any platform out of the water!
Edited 2010-03-12 09:42 UTC
Oh for crying out loud – talk about people taking extremes of POV.
Kishe: you’re comments are so optimistically thinking that there’s not even remotely based on reality.
1/ who’s going to tell all the gamers to switch platforms once they’ve got their gaming machine set up? I certainly can’t see them formatting their machines any time soon.
2/ who’s going to tell hardware manufacturers that they need to invest time getting all of their hardware drivers up to scratch. Sure, they’re probably about 75% there, but in gaming FPS matters and that 25% loss of advanced chip set functionality would end up being a deal breaker for many gamers who spent a small fortune on hardware.
3/ who’s going to tell the games developers that they need to re-program all of their current games for a completely different platform for free? (they’re not going to get much income for it as you’re talking about 50% of an existing Windows market share plus whatever casual gamers Linux has).
4/ who’s going to tell all of the above that they’ve only got 4 week in which to perform the migration?
Utter madness.
deathshadow: you’re not exactly being level-headed either.
Though you’re comments have much more basis on facts, you’re still somewhat overstating the reality.
1/ Video card drivers are easy to set up in Linux. In fact, even in Slackware I had little troubles.
2/ re:dual head graphics cards: How many Windows gamers use multiple monitors in games? Very very few is the answer. So why even bring up twin-views in a thread about gaming? All you’re doing is attacking a platform for the sake of attacking it – so how about we concentrate on the real issues instead?
The reality for OpenGL is somewhere inbetween both your comments.
I wish I could vote you up 100 times. I hate it when people take such extreme positions and bend facts to their personal agenda…
Don’t forget “and viable video drivers that were easy to configure and delivered 100% of the hardware performance, instead of 60% and having to turn off all the ‘extras’ if you want to run more than one display.
I VERY rarely agree with deathshadow as he is just a plain old bitter troll trying to bash Linux whenenver possible using any dirty arguments and “facts” taken out of his arse.
But here… well, I just installed Linux back cos I wanted to try something. I had also obtained a second monitor just recently and thought to try it out. Turned out that whenever I tried to turn on the second monitor in anything other than clone mode I got an error about the resolution being too high. I could get them both working if I set it to 640×480, but that’s just unacceptable. After some time messing around and googling I found out I had to manually add “Virtual” “2560×1024” to xorg.conf to get it working properly!
And then… when I had both monitors working fine I found out my graphics card doesn’t support shader programs under Linux and lacks several other features that all work properly under Windows. And that if I enable Compiz (ie. 3D desktop effects) it will leave part of my freaking desktop undrawn because the width of it is over 2048…
My graphics card isn’t the latest and greatest so I assume that’s why there are still features missing, but don’t people usually make the claim that open-source drivers are the best one can have and all that? And how about multi-monitor setups? Those are not uncommon and it’s still this flaky..
The propitiatory drivers are much much better than the open source ones when it comes to graphics cards.
Which makes sense when you think about it as you’d expect manufacturers drivers to be better than Microsoft’s own graphics drivers in Windows as well.
You’d expect the chip set makers to know the hardware better than the guys reverse engineering it.
50%?? Really? all those businesses going to give up Windows. All those grandmothers and students and truckdrivers are going to switch just because of games?
I highly doubt it would happen, even if Linux gained a DirectX-like API tomorrow.
See there still wouldn’t be any games for it – yet.
Indeed, and if my grandma had had balls we would have called her grandpa.
I love and use linux on a daily basis (work and home). But let’s be clear, it still has a lot of weak points. And those are weak points because we are comparing it to Windows. Which is unfair because neither is really supposed to be the replacement for the other since they cater to different needs/wants.
It is not just DirectX-like facilities which are needed if Linux was to compete directly with Windows (in the desktop at least). I would just settle for an actual estable interface both at the kernel-module and library level. Linux is great as a development plaform, but as a deployment for 3rd party commercial stand alone software-based solutions… it is a complete night mare: There is no standarization among distros (honestly, how many times does the wheel need to be reinvented), there is little stability in the interfaces (changes in the gnu libraries and kernel interfaces are “fun”), there is no standard user interface/experience (again how many times does it need to be reinvented and fractured?), there is no proper way to abstract the intrinsics from novice users or non-technically oriented/interested users, etc.
I know that has been discussed ad infinitum. But offering DirectX-like capabilities in Linux would not be a magic bullet, simply because there is a plethora of other hurdles which still would have to be addressed.
That being said, I am perfectly happy with linux and Windows targeting different spaces. The fragmentation which makes linux such a crappy platform for mass adoption, makes it an ideal development/experimentation platform.
There’s an easy solution to the fragmentation: Target Ubuntu.
It is the distro used by the vast majority of new users, and users who want something that “just works”.
I’ve got Ubuntu on mum’s netbook, and she loves it.
Really, the underlying OS is almost identical in most instances anyway, so creating a *.deb for ubuntu+other debian derivatives with a repo in it for updates, running alien on it for RPM systems, and then giving a tarball hits almost all users.
The main complain of OpenGL was the arcane API, is that fixed now?
Edited 2010-03-12 17:02 UTC
Depends on what you mean by “antiquated,” but I think the answer is, “not yet, but they’re working on it.” They’ve deprecated all the old cruft, and moved it out of the core specification into an optional backwards compatibility layer. Which is progress.
Actually, I think I’ve never heard of OpenGL 3.x. Is it that the Khronos Group won’t release an API until it’s deprecated or something ?
If you look at the dates you can see that there is a lot of active OpenGL development:
OpenGL 2.1 – July 2006
OpenGL 3.0 – July 2008
OpenGL 3.1 – May 2009
OpenGL 3.2 – August 2009, updated December 2009
OpenGL 3.3 – March 2010
OpenGL 4.0 – March 2010
Okay, I checked, it’s just either that Mesa have completely outdated OpenGL support or that the wikipedia fact about each release of OpenGL not being released in public domain until a new release is out is true.