AMD & NV drivers - now with OpenGL 4.2 & OpenCL 1.1 support

kick · Post by **kick** » Tue Mar 03, 2009 8:18 am

Great news (Direct2D support!)

kick · Post by **kick** » Thu Mar 19, 2009 9:25 am

ATI Catalyst 9.3 is available for download:

- 3 new GL extensions
- Full WDDM 1.1 support for Windows 7
- OpenGL works under Windows 7

- All problems with video acceleration are fixed

h4tred · Post by **h4tred** » Fri Mar 20, 2009 3:03 am

Its such a shame though that now there is little point to all these OpenGL improvements.

From a developer's point of view: OpenGL 3.0 is utter shit. DX9/10 completely whips OpenGL's ass.

I'm serious. I found actual reasons why DX > OGL. My world is shattered. I'm actually glad there is no such things as extensions in DX9/DX10. No need to mess around with shit like GLEW/GLUT/GLee or do crazy function pointer hacks

. Everything works like it should and does so equally. None of these fucked up shader compilers in video drivers for GLSL for certain cards.

And....a trimmed down Vista x64 isn't actually half bad. Its quite decent.

funkyass · Post by **funkyass** » Fri Mar 20, 2009 7:25 am

wait till you try win 7.

that being said, the OGL app I use has serious issues with the latest drivers.

go ATI go.

Dullaron · Post by **Dullaron** » Fri Mar 20, 2009 9:05 am

I don't use ATI cards anymore. Because of Ubuntu ATI driver set up. I never have a problem on Nvidia at all.

Oh yea I gave those ATI cards to my friend.

adventure_of_link · Post by **adventure_of_link** » Fri Mar 20, 2009 8:16 pm

Dullaron wrote:I don't use ATI cards anymore. Because of Ubuntu ATI driver set up. I never have a problem on Nvidia at all.

Oh yea I gave those ATI cards to my friend.

Not Ubuntu's fault, sorry. It's Linux in general. one reason why I tend to prefer nvidia to ATi.

the other, is that IIRC there's little display hacks you have to use in order to get things to display right on ATi cards. I believe sinamas ended up having to use one in gambatte.

h4tred · Post by **h4tred** » Sun Mar 22, 2009 10:24 pm

Oh yeah, tell me about it >_<

With my project, we had to use heaps of texture format hacks since ATI cards have a 50% slowdown with those texture formats. Not to mention the shader output being fucked in some situations >.>

Needless to say, ATI's reputation with OpenGL support is well deserved.

kick · Post by **kick** » Tue Mar 24, 2009 9:49 pm

Now with OpenGL 3.1 support!

Nine months ago the Khronos Group released the specification to OpenGL 3.0. OpenGL 3.0 brought version 1.30 of the GL Shading Language, the introduction of Vertex Array Objects, texture arrays, more flexible frame-buffer objects, and a number of other graphical features. What OpenGL 3.0 didn't bring was a major API revision that many developers had expected, and it was also arrived many months late. Today though, Khronos has announced OpenGL 3.1.

The OpenGL 3.1 specification also updates the GL Shading Language to GLSL version 1.40. OpenGL 3.1 introduces texture buffer objects, signed normalized textures, uniform buffer objects, more samplers, primitive restart, instancing, and a CopyBuffer API. In particular, the CopyBuffer API will be important when sharing buffers between OpenGL and OpenCL. In regards to OpenCL, Khronos also let loose that OpenCL implementations should begin appearing quite soon.

The OpenGL 3.1 specification can be downloaded on the Khronos web-site. The press release announcing the OpenGL 3.1 specification release can be found in their news room.

ATI and nVidia already have partial support for OGL 3.1 in their Windows and Linux drivers.

OpenGL 3.1 = awesome !

Squall_Leonhart · Post by **Squall_Leonhart** » Wed Mar 25, 2009 8:12 am

Opengl 3.1 is what 3.0 should've been?

kick · Post by **kick** » Wed Mar 25, 2009 8:18 am

Squall_Leonhart wrote:Opengl 3.1 is what 3.0 should've been?

Nope. OpenGL 3.2 will be the real deal (geometry shaders,tessellation and double-precision operations [with OpenCL])

franpa · Post by **franpa** » Wed Mar 25, 2009 10:04 am

http://developer.nvidia.com/object/opengl_3_driver.html

OpenGL 3.1 driver. (182.47)

kick · Post by **kick** » Wed Mar 25, 2009 4:07 pm

The new OpenGL 3.1 specification removed those features that were marked as deprecated in OpenGL 3.0.
A new 'ARB_compatibility' extension has been created to support the deprecated functionality.This single extension encapsulates all the removed functionality, and re-introduces that back into core OpenGL 3.1.This extension is not mandatory and NVIDIA, ATI, S3 can decide to add or not add it in their drivers.

kick · Post by **kick** » Wed Mar 25, 2009 6:10 pm

franpa wrote:http://developer.nvidia.com/object/opengl_3_driver.html

OpenGL 3.1 driver. (182.47)

Nice,but that driver still has partial support for OpenGL 3.1 - half of the features are not yet implemented.

It's just another marketing stunt by NVIDIA

Squall_Leonhart · Post by **Squall_Leonhart** » Wed Mar 25, 2009 6:23 pm

Nvidia has decided to continue to support any extension that is marked as deprecated and will continue to fix issues with them.

Deathlike2 · Post by **Deathlike2** » Thu Mar 26, 2009 12:33 am

Despite this thread, I'm still convinced that ATI still has to get its OpenGL shit together, particularly in Linux before I have any confidence in their OpenGL support.

These discussions tend to be a "hey, ATI's fixed their OpenGL e-penis except where it matters most, for developers".

h4tred · Post by **h4tred** » Fri Mar 27, 2009 4:31 am

Nope. OpenGL 3.2 will be the real deal (geometry shaders and double-precision operations FTW!)

..............

So, OpenGL is the "hey we don't innovate, we imitate!" API?

Seriously, my faith in D3D is growing more and more each day, and I am now in the stage of learning D3D based graphics programming. Its becoming more and more obvious that OGL is dead.

These discussions tend to be a "hey, ATI's fixed their OpenGL e-penis except where it matters most, for developers".

Yes and it shows how Khronos is trying thier hardest to bring thier specs in line with D3D. It should have been there from the start. All we are getting is empty promises and nothing but a lame imitation.

texture buffer objects

Just when D3D has this render target functionality perfect for so long, OGL is still trying to get it right.

.

We had:

* glReadPixels/Buffer
* PBuffers
* Framebuffer objects
* Pixel buffer objects

And now this??????????? Need I say more?

Deathlike2 · Post by **Deathlike2** » Fri Mar 27, 2009 2:14 pm

h4tred wrote:So, OpenGL is the "hey we don't innovate, we imitate!" API?

That was years ago.

Seriously, my faith in D3D is growing more and more each day, and I am now in the stage of learning D3D based graphics programming. Its becoming more and more obvious that OGL is dead.

It's not dead... just castrated because they are waiting for MS, instead of doing stuff on their own. Reactive vs Proactive...

Yes and it shows how Khronos is trying thier hardest to bring thier specs in line with D3D. It should have been there from the start. All we are getting is empty promises and nothing but a lame imitation.

It sounds more like too much arguing over details instead of trying to get it right for the good of OpenGL.

kick · Post by **kick** » Fri Mar 27, 2009 3:09 pm

New version of GLView with OpenGL 3.1 detection and GL3 bugfixes:
ftp://ftp2.realtech-vr.com/realtechv/pub/glview312.exe

Try it with the latest driver for your videocard and post the results

kick · Post by **kick** » Fri Mar 27, 2009 4:26 pm

With the 'old' ATI Catalyst 9.3 driver, GLView reports 60% of OpenGL 3.1 supported.

GLView's detection is fixed now. It shows full GL3.0 / GLSL 1.30 support if you select the GL3 context option.

kick · Post by **kick** » Fri Mar 27, 2009 7:29 pm

Deathlike2 wrote:Despite this thread, I'm still convinced that ATI still has to get its OpenGL shit together, particularly in Linux before I have any confidence in their OpenGL support.

http://www.phoronix.com/scan.php?page=a ... site&num=1

Nice!

Deathlike2 · Post by **Deathlike2** » Fri Mar 27, 2009 7:52 pm

That tells me nothing serious of consequence. I'd rather hear from people who have constantly whined about ATI's Linux drivers to report something positive.

kick · Post by **kick** » Fri Mar 27, 2009 9:39 pm

That tells me nothing serious of consequence. I'd rather hear from people who have constantly whined about ATI's Linux drivers to report something positive.

Catalyst 9.3 thread -> http://www.phoronix.com/forums/showthread.php?t=16220

Catalyst 9.4 (for Ubuntu) thread -> http://www.phoronix.com/forums/showthread.php?t=16059

The good thing is there are AMD developers and PR guys posting in those threads,so you can bitch and post your suggestions for improvement.Unlike the NVIDIA forums,the guys here listen to your complaints and report the issues to the driver developers

h4tred · Post by **h4tred** » Mon Mar 30, 2009 2:57 am

The good thing is there are AMD developers and PR guys posting in those threads,so you can bitch and post your suggestions for improvement.Unlike the NVIDIA forums,the guys here listen to your complaints and report the issues to the driver developers

And why are they doing it so late? These issues are meant to be fixed ages ago

I.S.T. · Post by **I.S.T.** » Mon Mar 30, 2009 4:41 am

Simple: All they're doing when on forums is PR. Nothing more, nothing less.

Squall_Leonhart · Post by **Squall_Leonhart** » Mon Mar 30, 2009 2:49 pm

kick wrote:
That tells me nothing serious of consequence. I'd rather hear from people who have constantly whined about ATI's Linux drivers to report something positive.
Catalyst 9.3 thread -> http://www.phoronix.com/forums/showthread.php?t=16220

Catalyst 9.4 (for Ubuntu) thread -> http://www.phoronix.com/forums/showthread.php?t=16059

The good thing is there are AMD developers and PR guys posting in those threads,so you can bitch and post your suggestions for improvement.Unlike the NVIDIA forums,the guys here listen to your complaints and report the issues to the driver developers

Nvidia listens as well, when the people complaining aren't random retards. A number of bugs i've reported through the years have been fixed.