Google Gets Into The 3D Driver Game

[phoronix]

Phoronix: Google Gets Into The 3D Driver Game

To help out the adoption of WebGL, the Khronos-backed API originally started by Mozilla that seeks to let web developers tap into modern graphics processors via the web-browser natively, has caused Google to get into the graphics driver game. WebGL binds to OpenGL ES 2.0, and with the Microsoft graphics drivers being more DirectX-optimized rather than OpenGL, Google's playing to Microsoft. Google wants more users to be able to use WebGL, particularly when running the Chrome browser, so they have just announced the Almost Native Graphics Layer Engine. The objective of ANGLE is to just take the subset of the OpenGL ES API exposed by WebGL and to translate those extensions into their DirectX 9.0c equivalents...

http://www.phoronix.com/vr.php?view=ODA4MQ

[FireBurn]

Does anyone else think this is terrible news?

Instead of companies putting effort into their shoddy OpenGL implementations on windows and as a side effect improving their Linux binary drivers which shares the code, Google are allowing companies to use wrap WebGL to DirectX instead

This does nothing for the OpenGL cause at all!

[d2kx]

I don't exactly understand that move either, but what I know is:

Google likes the web
They like WebGL and want it to have success
Their new distribution Chrome OS is based on Linux and uses OpenGL
They want Chrome OS to be able to do all the things the web offers

I wouldn't panic

Also, with the recent push of Steam for Mac and regular new OpenGL releases, I think the driver's OpenGL implementation will improve anyway.

[Remco]

I'm pretty sure AMD and NVIDIA won't be abandoning their OpenGL stack. The scientific industry wants OpenGL. Intel may depend on ANGLE, but I have a hard time believing they would implement anything beyond WebGL anyway.

The other side of this is that ANGLE makes WebGL feasible for all computers. WebGL has the greatest chance of making OpenGL relevant on Windows again, and if 50% of the computers can't even run WebGL, then it will fail.

[Jimmy]

It would be nice to hear from the ANGLE team what aspects of OpenGL are so poorly supported on Windows that reimplementing OpenGL support is necessary.

If ATI and NVidia share most of their implementation across platforms it would seem to follow that OpenGL support on Windows isn't that bad for ATI or NVidia based solutions. Is it that OpenGL sucks that bad for other vendors?

It just seems odd that they would need to take this approach at all.

[Remco]

It would be nice to hear from the ANGLE team what aspects of OpenGL are so poorly supported on Windows that reimplementing OpenGL support is necessary.

If ATI and NVidia share most of their implementation across platforms it would seem to follow that OpenGL support on Windows isn't that bad for ATI or NVidia based solutions. Is it that OpenGL sucks that bad for other vendors?

It just seems odd that they would need to take this approach at all.

Intel has 50% of the GPU market. And Intel's OpenGL is crap.

[Remco]

Plus, Windows does not have OpenGL drivers for NVIDIA or AMD by default.. To get OpenGL at all, you need to download NVIDIA's or AMD's updated drivers manually.

[Craig73]

(like was the case with ATI Catalyst and Lucid, until yesterday)

Lucid - that's in Beta right? (ie, How does this add to the conversation?)

[dfx.]

google scaring me more and more each day with more of proprietary crap - maybe it turns evil after all...

[bridgman]

This is really only useful on Windows, isn't it ? If so, the downside seems limited and this does seem like a more OpenGL-supportive approach than implementing native DX support in WebGL. Based on that, I don't really see a problem with this.

One question is what happens when running on a system which has both OpenGL and DirectX support. Presumably there will be an option to choose the graphics stack ?