Jump to content

  • Log In with Google      Sign In   
  • Create Account

Interested in a FREE copy of HTML5 game maker Construct 2?

We'll be giving away three Personal Edition licences in next Tuesday's GDNet Direct email newsletter!

Sign up from the right-hand sidebar on our homepage and read Tuesday's newsletter for details!


We're also offering banner ads on our site from just $5! 1. Details HERE. 2. GDNet+ Subscriptions HERE. 3. Ad upload HERE.


OpenGL 4.4 spec is published


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
22 replies to this topic

#1 TheChubu   Crossbones+   -  Reputation: 4555

Like
4Likes
Like

Posted 24 July 2013 - 09:40 AM

Well, since I didn't saw any thread about it, here it is:

 

OpenGL 4.4 released.

 

OpenGL registry where you can see the pdfs of the spec.

 

And nVidia's drivers for it (Linux/Windows/x86/x86_64/yaddayadda). There are a few new extensions also available for OpenGL 3 hardware.


Edited by TheChubu, 24 July 2013 - 09:43 AM.

"I AM ZE EMPRAH OPENGL 3.3 THE CORE, I DEMAND FROM THEE ZE SHADERZ AND MATRIXEZ"

 

My journals: dustArtemis ECS framework and Making a Terrain Generator


Sponsor:

#2 Koehler   Members   -  Reputation: 228

Like
1Likes
Like

Posted 24 July 2013 - 10:19 AM

I'm excited about ARB_Sparse_Texture, though I'm a little confused as to why they don't support any of the 3-component texture formats.



#3 Promit   Moderators   -  Reputation: 7335

Like
2Likes
Like

Posted 24 July 2013 - 10:53 AM

Not a very interesting revision, but the new extensions (sparse texture, bindless texture) are intriguing...



#4 MarekKnows.com   Members   -  Reputation: 666

Like
0Likes
Like

Posted 24 July 2013 - 11:06 AM

Well, since I didn't saw any thread about it, here it is:

 

OpenGL 4.4 released.

 

OpenGL registry where you can see the pdfs of the spec.

 

And nVidia's drivers for it (Linux/Windows/x86/x86_64/yaddayadda). There are a few new extensions also available for OpenGL 3 hardware.

 

Thanks for the heads up.  I'll have to check it out.


Edited by MarekKnows.com, 24 July 2013 - 11:07 AM.

---
Free C++, OpenGL, and Game Development Video Tutorials @
www.MarekKnows.com
Play my free games: Ghost Toast, Zing, Jewel Thief


#5 Makers_F   Members   -  Reputation: 793

Like
1Likes
Like

Posted 24 July 2013 - 03:04 PM

I think the biggest improvement is the conformace test. From what i head from who prefer DX over GL is that the drivers sometimes have different behaviours for different cards (with openGL). With this change, all the driver will(?should?) have the same behaviour, making it easier to develop openGL programs.



#6 Chris_F   Members   -  Reputation: 2435

Like
0Likes
Like

Posted 24 July 2013 - 08:49 PM

So, OpenGL 4.4 specs have been published, and AMD still doesn't have a working implementation of OpenGL 4.3. This is wonderful. rolleyes.gif



#7 Hodgman   Moderators   -  Reputation: 30958

Like
2Likes
Like

Posted 24 July 2013 - 08:52 PM

I'm excited about ARB_Sparse_Texture, though I'm a little confused as to why they don't support any of the 3-component texture formats.

GPU hardware hasn't supported 3-component texture formats for a long time (aside from packed formats like DXT1).

If you ask GL to give you an RGB texture, on the GPU it will allocate an RGBA texture and pretend that the alpha channel doesn't exit...

I think the biggest improvement is the conformace test. From what i head from who prefer DX over GL is that the drivers sometimes have different behaviours for different cards (with openGL). With this change, all the driver will(?should?) have the same behaviour, making it easier to develop openGL programs.

Yeah that's something that I always have a whinge about, so this makes me very happy biggrin.png

[the ARB] has created the first set of formal OpenGL conformance tests since OpenGL 2.0 [and] full certification is mandatory for OpenGL 4.4 and onwards


Edited by Hodgman, 24 July 2013 - 08:54 PM.


#8 Sik_the_hedgehog   Crossbones+   -  Reputation: 1811

Like
0Likes
Like

Posted 24 July 2013 - 09:57 PM

Not gonna comment on the features since I'm still trying to play catch up (this is what happens when you live in a place where any upgrade to a PC ends up costing an entire salary, let alone getting a new computer that isn't already outdated).

 

I'm curious about the certification part, though. Yeah, sure, having a guarantee that drivers always work the same is nice, but how do they plan to enforce it? Does this mean it'll be outright illegal to release an OpenGL driver that isn't certified? I can see that being a massive issue for FOSS drivers (which does matter on Linux). Does anybody have exact details on what certification allows?

 

EDIT: should have checked more carefully

http://www.khronos.org/conformance/

 

OK, it's mostly a trademark issue (so e.g. Mesa probably still would be safe since it doesn't call itself OpenGL). It seems that FOSS implementations still would be able to go through the implementers program if they want to use the name (not the adopters one due to the fee).


Edited by Sik_the_hedgehog, 24 July 2013 - 10:02 PM.

Don't pay much attention to "the hedgehog" in my nick, it's just because "Sik" was already taken =/ By the way, Sik is pronounced like seek, not like sick.

#9 samoth   Crossbones+   -  Reputation: 4913

Like
0Likes
Like

Posted 25 July 2013 - 03:18 AM

So, OpenGL 4.4 specs have been published, and AMD still doesn't have a working implementation of OpenGL 4.3. This is wonderful. rolleyes.gif

Well, AMD is much less of an issue than Intel, they (AMD) are actually pretty good nowadays. Intel is the real problem. They not only do not have a working implementation of OpenGL 4.0, but they also do not have a working implementation of OpenGL 3.0.

 

Which makes me wonder what this entire certification thing will be good for at all.

 

Intel will not pull functional 3.0 (let's not even imagine 4.4) drivers out of their magic hat, but Intel integrated GPUs are the main GPU in every El Cheapo computer, and in the major share of non-tablet computers anyway, too. And, outside the world of Android, they're pretty much omni-present in tablets as well.

 

Which will probably mean no more and no less than OpenGL will simply not be supported (or supported even worse as it is now) on a considerable share of hardware. Sorry for being pessimistic, but I just can't see Intel producing a quality 4.x driver and undergo certification any time soon. They'll just show everyone the middle finger, knowing their CPUs are sold anyway.



#10 Chris_F   Members   -  Reputation: 2435

Like
0Likes
Like

Posted 25 July 2013 - 03:45 AM

 

So, OpenGL 4.4 specs have been published, and AMD still doesn't have a working implementation of OpenGL 4.3. This is wonderful. rolleyes.gif

Well, AMD is much less of an issue than Intel, they (AMD) are actually pretty good nowadays. Intel is the real problem. They not only do not have a working implementation of OpenGL 4.0, but they also do not have a working implementation of OpenGL 3.0.

 

Which makes me wonder what this entire certification thing will be good for at all.

 

Intel will not pull functional 3.0 (let's not even imagine 4.4) drivers out of their magic hat, but Intel integrated GPUs are the main GPU in every El Cheapo computer, and in the major share of non-tablet computers anyway, too. And, outside the world of Android, they're pretty much omni-present in tablets as well.

 

Which will probably mean no more and no less than OpenGL will simply not be supported (or supported even worse as it is now) on a considerable share of hardware. Sorry for being pessimistic, but I just can't see Intel producing a quality 4.x driver and undergo certification any time soon. They'll just show everyone the middle finger, knowing their CPUs are sold anyway.

 

 

 I really don't consider Intel to be that big of any issue. Their integrated graphics are in a completely different class compared to AMD and Nvidia's dedicated GPUs. I mean whats the real advantage to being able to enable the latest OpenGL 4 / DX11 level features in a game if its going to run at 5fps?


Edited by Chris_F, 25 July 2013 - 03:46 AM.


#11 Erik Rufelt   Crossbones+   -  Reputation: 3516

Like
0Likes
Like

Posted 25 July 2013 - 04:38 AM

 

Intel will not pull functional 3.0 (let's not even imagine 4.4) drivers out of their magic hat, but Intel integrated GPUs are the main GPU in every El Cheapo computer, and in the major share of non-tablet computers anyway, too. And, outside the world of Android, they're pretty much omni-present in tablets as well.

 

Which will probably mean no more and no less than OpenGL will simply not be supported (or supported even worse as it is now) on a considerable share of hardware. Sorry for being pessimistic, but I just can't see Intel producing a quality 4.x driver and undergo certification any time soon. They'll just show everyone the middle finger, knowing their CPUs are sold anyway.

 

 

I would be surprised if they added this certification if Intel haven't already said yes to it. What would be the point if it's still just AMD/Nvidia?

 

 

 

 I really don't consider Intel to be that big of any issue. Their integrated graphics are in a completely different class compared to AMD and Nvidia's dedicated GPUs. I mean whats the real advantage to being able to enable the latest OpenGL 4 / DX11 level features in a game if its going to run at 5fps?

 

 

That seems rather irrelevant in the context of conformance, where the point is that any features should behave the same, which is just as important if one only uses 2.0 functionality. AMD/Nvidia are already close enough, and the real advantage of the conformance tests would be when writing an application that doesn't require the beefiest hardware and being able to rely on it working the intended way on any device.



#12 mhagain   Crossbones+   -  Reputation: 8136

Like
0Likes
Like

Posted 25 July 2013 - 06:27 AM

Certification is just required for GL4.4+, so all that Intel (or AMD for that matter, should they be so inclined) have to do is freeze their implementation at a pre-4.4 level and hey-presto!  No need for certification and they can continue to ship driver bugs.

 

I wouldn't underestimate Intel, by the way.  Haswell is looking pretty good, is beating comparable parts from AMD, and in a couple more generations we may well see them emerging as a third serious player in the market.


It appears that the gentleman thought C++ was extremely difficult and he was overjoyed that the machine was absorbing it; he understood that good C++ is difficult but the best C++ is well-nigh unintelligible.


#13 phantom   Moderators   -  Reputation: 7395

Like
2Likes
Like

Posted 25 July 2013 - 06:40 AM

I wouldn't underestimate Intel, by the way.  Haswell is looking pretty good, is beating comparable parts from AMD, and in a couple more generations we may well see them emerging as a third serious player in the market.


Yeah, this.

Both in desktop and mobile the Intel machine has woken up and starting to push serious resources into the development; with AMD's CPU division sucking away the profits from graphics if Intel can keep up investment they could move into second place.

As for GL4.4; there isn't really a great deal to it.
From the headline features;

- Buffer Storage has mostly provoked arguements as to how useful it'll be (more so when the notes on the extension say that at least one of the bits might be ignored) - I'm pretty sure this also basically mimics D3D11's buffer controls

- Async Queries could be useful if you are doing anything which requires GPU output which would normally bounce thru a CPU buffer

- Shader Variable Layout, while intresting from a 'yay!' point of view is again basically a HLSL parity feature

- Multi-bind is a good addition but nothing earth shattering (and it's bizzare it wasn't about before... see D3D10)

- The 10-11-11 vertex format support is just... well, sane.. again, surprising it wasn't there before.

There are some intresting extensions about (sparse, bindless, draw parameters, variable group size(!), indirect parameters) but the core feels like a 'tidying up missing features vs D3D11' really.

Maybe GL4.5 will bring something new to the table in the core, we'll see, as D3D is basically stuck at D3D11 due to the enforcement of Win8.1 for D3D11.2 - OGL has basically a 3rd chance to try and become a viable option again.

#14 Koehler   Members   -  Reputation: 228

Like
0Likes
Like

Posted 25 July 2013 - 11:41 AM

GPU hardware hasn't supported 3-component texture formats for a long time (aside from packed formats like DXT1).


If you ask GL to give you an RGB texture, on the GPU it will allocate an RGBA texture and pretend that the alpha channel doesn't exit...
 

 

Learn something new every day! Good to know this.. 



#15 TheChubu   Crossbones+   -  Reputation: 4555

Like
0Likes
Like

Posted 25 July 2013 - 08:32 PM

So, OpenGL 4.4 specs have been published, and AMD still doesn't have a working implementation of OpenGL 4.3. This is wonderful. rolleyes.gif

As far as I know, the latest driver release should fully support OpenGL 4.3.


"I AM ZE EMPRAH OPENGL 3.3 THE CORE, I DEMAND FROM THEE ZE SHADERZ AND MATRIXEZ"

 

My journals: dustArtemis ECS framework and Making a Terrain Generator


#16 Chris_F   Members   -  Reputation: 2435

Like
2Likes
Like

Posted 25 July 2013 - 08:40 PM

 

So, OpenGL 4.4 specs have been published, and AMD still doesn't have a working implementation of OpenGL 4.3. This is wonderful. rolleyes.gif

As far as I know, the latest driver release should fully support OpenGL 4.3.

 

 

Yes, it fully supports OpenGL 4.3, except when it crashes due to bugs.



#17 TheChubu   Crossbones+   -  Reputation: 4555

Like
0Likes
Like

Posted 27 July 2013 - 08:58 PM

Yes, it fully supports OpenGL 4.3, except when it crashes due to bugs.

Don't be like that, its getting there! :P


"I AM ZE EMPRAH OPENGL 3.3 THE CORE, I DEMAND FROM THEE ZE SHADERZ AND MATRIXEZ"

 

My journals: dustArtemis ECS framework and Making a Terrain Generator


#18 mhagain   Crossbones+   -  Reputation: 8136

Like
0Likes
Like

Posted 27 July 2013 - 09:49 PM

Don't be like that, its getting there! tongue.png

 

"Getting there" isn't good enough.  GL4.3 has been specified for the past year, AMD are a member of the standards body that specified it, they should have had a full GL4.3 driver long ago.


It appears that the gentleman thought C++ was extremely difficult and he was overjoyed that the machine was absorbing it; he understood that good C++ is difficult but the best C++ is well-nigh unintelligible.


#19 Ingenu   Members   -  Reputation: 912

Like
0Likes
Like

Posted 28 July 2013 - 03:23 AM

They should put more effort into streamlining the API again, and I'm not sure why we don't have a single state vector yet, neither why samplers are not solely shader side for example.

It would also be nice to be able to generate and submit a command buffer sequence easily. (Create it once, resubmit it as many times as you want, sorta like a copy/paste operation.)

 

How many ways do we need to specify a typed multi-dimensional array ?

BufferData, BufferStorage, TextureImage*, TexStorage*... That's way too many...

 

The API is at last giving us access to hardware features that have been available for years now, but the API isn't making the jump I'd like to see and change once and for all into OpenGL Lean & Mean, a promise from 2002 for OpenGL 2.0... (Long before 3.0/Long Peaks)


Edited by Ingenu, 28 July 2013 - 03:25 AM.

-* So many things to do, so little time to spend. *-

#20 samoth   Crossbones+   -  Reputation: 4913

Like
0Likes
Like

Posted 28 July 2013 - 05:03 AM

What I don't get about query buffers is that for most query objects (including occlusion and timer, the most interesting ones) it says that at most one query can be active at a time.

 

In other words, you can now read many queries into a buffer object to avoid stalls and to avoid a round-trip to the CPU, but you can still only run one query at a time. Which, frankly, isn't so much different.

 

The only useful application is really transform feedback and/or geometry shader (where more than one query at a time can be active).






Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS