Jump to content

  • Log In with Google      Sign In   
  • Create Account


Matias Goldberg

Member Since 02 Jul 2006
Offline Last Active Today, 05:35 PM

Posts I've Made

In Topic: HLSL load() of directx 10 in directx 9

25 August 2014 - 04:47 PM

Fun AMD hack: If you set filtering to point, but mipmapping to anisotropic without mipmaps in the texture, and use sample2DLod, the driver will recognize it and you should get faster sampling of single pixels. Same with DX10's SampleLevel.

But profile it first; just to be sure it's done right.


In Topic: DXT normal map compression

22 August 2014 - 05:51 PM

You should read Understanding BCn Texture Compression Formats from Nathan Reed.

 

It should clear your questions. The only thing to add is that, before we had BC5 (which was only introduced in DX10; unless you used the ATI hacks); we used the technique you linked above: DXT5 and store the components in the green and alpha channel.

The alpha channel in DXT5 is encoded exactly the way BC5 does for its components. However the RGB components are not. They're encoded like in BC1, which delivers lower quality for our purposes. The Green channel is used over the Red and Blue because it has one more bit of precision.

 

This means BC5 beats DXT5 in terms of quality because the latter wastes precision on unused Red & Blue channels.


In Topic: Next Generation OpenGL

21 August 2014 - 12:00 PM

I'm gonna call the redesign as NextGL or NGL

Something that might be worth saying, a NextGL API that is low level means that it is perfectly possible to provide an efficient OpenGL 4.x (and older) implementation working on top of NextGL (i.e. a GL to NextGL wrapper).

 

Such implementation could be done by anyone (i.e. open source). No need for the driver developer to ever care about it again.

 

The same could be said about an OGL to DX12 wrapper; however NGL will be cross plattform too, and thus a GL to NGL wrapper can be much more reusable (i.e. Android, Linux & SteamOS, Windows; and if Apple joins; the iOS family too).


In Topic: Next Generation OpenGL

21 August 2014 - 10:10 AM

Despite the lack of confidence resulting from previous failures to deliver Long Peaks, this time looks like it will happen.

 

There is great people behind the GL (like Graham Sellersfrom AMD & Christophe Riccio). There is Mantle that can serve as a baseline (honestly if the just Find & Replace 'gr' for 'gl' we all would be happy!) GL has been adding a lot of so-much-needed functionality recently at an increasing rate (the Board used to be an insanely slow turtle!!!). They've added Johan Andersson who is very experienced and more importantly, results oriented.

 

Furthermore, there is a strong need for GL to move forward: 5 years ago everything was Windows and DirectX.

Today we have Google pushing GL into Android (and a countermeasure against Apple's Metal API) with a huge market share, Qualcomm also wants to push GL in mobile; GL still being the only choice to target China's huge WinXP user base with DX10 & 11 GPUs; Valve wants to push GL in their SteamOS, and MS still pushing their latest DX updates tied to OS (not to mention that they fell asleep to develop DX12)

Basically, we are like in 2006 situation with DX10+Vista vs Long Peaks. Except this time there is more incentive, a Board that is actually acomplishing things, and a Mantle proof of concept.

 

I hope this time they make it.


In Topic: glDrawElements isn't drawing anything

13 August 2014 - 07:47 AM

You're using, GL_UNSIGNED_INT, m_elements; but it's common to use unsigned shorts (16-bit indices). Are you sure you didn't mean GL_UNSIGNED_SHORT?


PARTNERS