Steam on Linux: Return of OpenGL?

[Arkadrel]

I like there being a "open" alternative to mircosofts DirectX.
Even if mircosoft probably sits on like 90% of the pc gameing market, because of DirectX.

If anything this might give MS slightly more incentive to putting more R&D into DirectX, because without it, the causual OS users *could* opt to shift OS (I believe gameing plays a big part in window's success).

That said, I dont really see this helping Linux all that much.

But like others have said, on the mobil front, OpenGL is the standart everyone uses, so OpenGL isnt going anywhere anytime soon.

 

[Dark Shroud]

The big companies already don't put in the effort on the PC front.

I hope some of you don't expect them to actually start developing for Linux. Just because Valve will port their engine to Linux.

 

[Kalessian]

PS3 uses a proprietary API but has an openGL layer over the top in the case where devs were too lazy to learn to use the more powerful base API.

Not sure why there's so much automatic negativity following announcements such as this on these forums. DirectX isn't going anywhere, and if OpenGL gains traction, what do you care?

There are many games already ported to Linux, especially Indie games. Take a look at the money brought in from the *nix crowd on the humble indie bundles, it was probably well worth the effort. http://www.inatux.com/?r=humble-indie-bundle-shows-gnu-linux-gaming-statistics

Unigine, Unreal Engine, iD's engines, and now Source all have opengl render paths and have made commitments to supporting linux at some point in the past few years. If the functionality is there and Steam Linux can nab some extra cash for a dev I don't see why they wouldn't try it.

Edit: Serious Engine 3.5 has OpenGL path: http://forums.steampowered.com/forums/showthread.php?t=2560568 linux port soon. Also, don't forget blizzard games with their OpenGL (WoW has >10 mill players)

 

[Olikan]

There are many games already ported to Linux, especially Indie games. Take a look at the money brought in from the *nix crowd on the humble indie bundles, it was probably well worth the effort. http://www.inatux.com/?r=humble-indie-bundle-shows-gnu-linux-gaming-statistics

OMG! Valve must be very, very rich

 

[Kalessian]

sigh... is picking on indie devs supposed to be some kind of joke

EDIT: Valve dedicated to improving the (graphics) ecosystem of linux- http://www.phoronix.com/scan.php?page=news_item&px=MTE0MTY

 

[mv2devnull]

The big companies already don't put in the effort on the PC front.

The big game companies ...

There is, you know, more in life than just games.

 

[2is]

The big game companies ...

There is, you know, more in life than just games.

There are also other threads and other forums that aren't talking about gaming. Since this thread is, maybe you can go post somewhere else and have some relevance.

 

[Mopetar]

I've always been very interested in OpenGL and preferred its use over DirectX, for obvious reasons. Portability and platform independence being front and center. With *Steam coming to Linux in 2012 according to Gabe Newell, I would assume OpenGL will be making a comeback.

No, Linux is such a small market that it's not going to amount to anything in the grand scheme of things. That and I think that we're past the point of the Linux desktop ever taking off. By the time it manages to gain any significant market share, the hay day of the desktop as the predominant computing platform will be over and most people will be using smart phones, tablets, or their successors.

If anything helps OpenGL overtake DirectX it will be Sony/Nintendo consoles, mobile OS (iOS/Android) games using OpenGL ES, and possibly Mac OS X slowly eating into Windows sales. Even then it's going to be a pretty slow transition if it happens at all.

Linux support on Steam isn't going to do anything for helping convincing people to use OpenGL.

 

[Obsoleet]

Linux support on Steam isn't going to do anything for helping convincing people to use OpenGL.

As someone who works in the game development industry, I would disagree. Management at my company has stated that making the most portable product is paramount.
We all have seen what comeback Apple has made in the last few years. If anyone besides Microsoft takes over, which there's plenty of competition (Google, Apple, Linux)- it's going to be OpenGL based. Not to mention the mobile and console scenes, which are more important than the desktop gaming scene- are already on OpenGL.

So you can put your eggs into 1 basket (MS DX), or you can go with the future, which is mobile/console/alternativeOSs. That's not to say that you can't use OGL on Windows, so you're not missing out on any market.

Having a good product is more important, and the singlemost important desktop gaming OS is Windows, but OGL works on that too, so I'm not sure why the negative reaction to OGL gaining traction (which Steam on Linux will definitely help with).

 

[Mars999]

I code in OpenGL and have done DX, I like OpenGL better IMO and love the idea of code once run anywhere vs. locked down....

For all who don't know GL4.2 actually ahead of DX11 in terms of better gfx options and with bindless states GL should be faster than DX.

A few years ago I did some terrain rendering in DX9 and Fuc* U DIP() is so SLOW, vs. GL's calls a lot faster like 2x at least....

So to all you GL haters take a look around and just get it through you head it doesn't matter to you what API coders like Obsoleet and I use as long as you get better gfx and can run it on Linux(thank GOD finally Steam), OSX, Win32 you should not care.

IMO have two API's is killing progress and of HUGE waste of cash for AMD/Nvidia just drop DX and save some money and use GL IMO....

Also IMO in a few years API's probably will be gone anyway so much power showing up software based rendering or ray tracing with some rasterization is going to be looked at more than it was before.

 

[Anarchist420]

Also IMO in a few years API's probably will be gone anyway so much power showing up software based rendering or ray tracing with some rasterization is going to be looked at more than it was before.

That's what I'm hoping for as well, but there are two things standing in the way of software based rendering... 1) patents and 2) the fact that intel seems hell-bent on wasting CPU die space with having on-chip tesselation, ROPs, and depth/stencil units. They should make a haswell with 8 cores and DX/GL compliant texture units (and an iTMDS) and simply emulate blending and depth through the AVX 2 units. Or if they didn't want to make 8 cores, then they can go with 4 cores and double or triple the number of AVX 2 units plus the texture units and i(ntegrated) TMDS. They could also use DDR4 (or go with 6 or 8 DDR3 channels) and bandwidth problems would be completely solved (DDR4) or reduced (more DDR3 channels). DDR3 and DDR4 have lower latency than GDDR5, so some of the lost bandwidth will be partially made up for by significantly lower latency.

 

[Mars999]

That's what I'm hoping for as well, but there are two things standing in the way of software based rendering... 1) patents and 2) the fact that intel seems hell-bent on wasting CPU die space with having on-chip tesselation, ROPs, and depth/stencil units. They should make a haswell with 8 cores and DX/GL compliant texture units (and an iTMDS) and simply emulate blending and depth through the AVX 2 units. Or if they didn't want to make 8 cores, then they can go with 4 cores and double or triple the number of AVX 2 units plus the texture units and i(ntegrated) TMDS. They could also use DDR4 (or go with 6 or 8 DDR3 channels) and bandwidth problems would be completely solved (DDR4) or reduced (more DDR3 channels). DDR3 and DDR4 have lower latency than GDDR5, so some of the lost bandwidth will be partially made up for by significantly lower latency.

Actually GDDR6 is going to be out about the same time DDR4 shows up... I hear it's going to be the last GDDR... so who knows after that....

I agree, bring on the many core CPU's I really hate the whole way you have to go about coding in API's like DX or GL to do shadows ect... seems like such a waste or inefficient methods draw, draw, check blah blah sigh....

Coolbeans someone else here agrees with me!

BTW AVX1/2 ROCK!!!!

Can't wait to see FMA instructions are going to be interesting.

 

[Subyman]

iOS uses OpenGL and OpenAL. OpenGL's fate lies in iOS IMO.

 

[Mars999]

iOS uses OpenGL and OpenAL. OpenGL's fate lies in iOS IMO.

Incorrect, iOS uses GLES which is a derivative of GL... just to clarify....

Also WebGL is taking off also! Thank GOD!

 

[f1sherman]

I'd love to see OpenGL retaking a foothold @PC.

I allways felt OpenGL games were more responsive. Like no input/display lag at all.

 

[blastingcap]

http://www.bit-tech.net/hardware/2012/07/18/build-your-own-steam-box/1

 

[Subyman]

Incorrect, iOS uses GLES which is a derivative of GL... just to clarify....

Also WebGL is taking off also! Thank GOD!

Its not incorrect, OpenGL ES is a subset of OpenGL, so OpenGL includes it

 

[Mars999]

Its not incorrect, OpenGL ES is a subset of OpenGL, so OpenGL includes it

No it doesn't, try running GLES code without drivers that have GL. You can't just use GLES code in GL

 

[ViRGE]

No it doesn't, try running GLES code without drivers that have GL. You can't just use GLES code in GL

Correct. OpenGL 4.1 was the first version of OpenGL to also be OpenGL ES 2.0 compliant.

 

[BFG10K]

IMO have two API's is killing progress and of HUGE waste of cash for AMD/Nvidia just drop DX and save some money and use GL IMO....

OpenGL’s progress would be almost completely stagnant if it wasn’t for DirectX marching along at such a brisk pace. DirectX also influences the new hardware features from GPU vendors. OpenGL just plays catch-up.

Furthermore, how does OpenGL plan on supporting other functions from DirectX, such as input and sound? DirectX is more than just graphics.

Also IMO in a few years API's probably will be gone anyway so much power showing up software based rendering or ray tracing with some rasterization is going to be looked at more than it was before.

Intel tried the same fantasy with Larrabee, and look how many billions it cost them. Software rendering will never be viable because no matter how many cores you throw at the problem, you’ll still be castrated by system memory bandwidth. You’ll also be castrated by the TDP limitations associated from limited CPU socket space.

You also can’t take away APIs from any modern OS using pre-emptive multi-tasking, dynamic memory allocation and hardware abstraction because such an OS can’t allow programs to directly take control of the hardware.

As for ray tracing, if that starts taking off you can bet GPU vendors will figure out how to add dedicated hardware units much faster than any generic CPU core.

 

[Pottuvoi]

Titles like Rage continually prove cutting edge games games can be done with OpenGL. Do you think this is a big win for OpenGL? Does this score a big win for idtech5 as well?

If developers get bashed for a driver problems like in case of Rage, it really doesn't paint good future for GL.

didn't playstation 3 use OpenGL?

There is implementation of OpenGL ES for Ps3, but it simply is not a viable option for game developers as it restricts the 3D engine severely.

This is what programmers think of GL on ps3..
'no, no one in their sane mind uses OpenGL on PS3' - Repi

What they use is LibGCM, which is close to metal api and was actually used to create the OpenGL path for Ps3.
There has been some talk about it and word is that it is very good.

 

[Shmee]

ETQW is still around and played btw.

 

[Mars999]

OpenGL’s progress would be almost completely stagnant if it wasn’t for DirectX marching along at such a brisk pace. DirectX also influences the new hardware features from GPU vendors. OpenGL just plays catch-up.

Furthermore, how does OpenGL plan on supporting other functions from DirectX, such as input and sound? DirectX is more than just graphics.


Intel tried the same fantasy with Larrabee, and look how many billions it cost them. Software rendering will never be viable because no matter how many cores you throw at the problem, you’ll still be castrated by system memory bandwidth. You’ll also be castrated by the TDP limitations associated from limited CPU socket space.

You also can’t take away APIs from any modern OS using pre-emptive multi-tasking, dynamic memory allocation and hardware abstraction because such an OS can’t allow programs to directly take control of the hardware.

As for ray tracing, if that starts taking off you can bet GPU vendors will figure out how to add dedicated hardware units much faster than any generic CPU core.

Actually incorrect, GL has extensions which allow features way before DX ever gets to dream about them. So no DX doesn't drive GL, it's the other way around.

As for the last statement correct, I agree, and could care less who makes an accelerator to do it, but as long as they move to that realm is all I care. I am sick of well you don't have feature "X" on you POS old Gfx card so no you can run this or that. With ray tracing or software you could run the game still, but it may run slower but at least you could run it, or turn down the resolution for cheap asses...

 

[Anarchist420]

As for ray tracing, if that starts taking off you can bet GPU vendors will figure out how to add dedicated hardware units much faster than any generic CPU core.

That wouldn't surprise me.

Intel tried the same fantasy with Larrabee, and look how many billions it cost them. Software rendering will never be viable because no matter how many cores you throw at the problem, you’ll still be castrated by system memory bandwidth. You’ll also be castrated by the TDP limitations associated from limited CPU socket space.

I'm not sure the memory bandwidth would be a problem with DDR4. In terms of performance/watt, I'd argue that the integrated ROPs and depth units (and onchip DRAM soon) are not worthit. Intel is only slacking because AMD either can't or won't make anything good.

Also, wasn't the most powerful Larrabee design deemed inadequate like 3 years ago just because it wasn't more than 60% of a GTX 280's performance? Intel's fabs are also even more ahead of the ones used by the GPU vendors than they were 3 years ago.

 

[clok1966]

Looking for APPLE to revive OpenGL is ... well wishfull thinking.. APPLE will not support anything that cant directly fill its pockets. OpenGL is free.. they may take some openGL and call it APPLE GL and licence the crap out of it and make it standard for anything sold on Istore.. but it wont be OpenGL Linux.. not untill it works as easy windows. sorry it just doesnt.. and doubtfully never will.

Apple stuff works fine.. but its closed system will never work. Dont get me wrong, it may build and overtake windows someday.. but thats more about people trying to be trendy (ipads and Iphones) the WORK related stuff.

There are some big changes comming.. Ipad may be the way of the future.. smaller harder to read screens (especialy if used outside on a "MOBILE" platform, kinda defeats the purpose, go inside so you can read it), keyboards that cover my apps so i cant see it.. dont fit in my pocket, slower then a real Laptop or PC.. WHy didnt i just take a laptop along again? oh ya.. its not trendy, its just usefull.. of course if you just need facebook and email its better, right? (me i just would use my phone, i already got it with me).