[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Public WebGL] Odd behavior on eePC netbook.



My son just bought an eePC netbook - a tiny WinXP laptop with an Intel
GMA950 graphics chip.  According to the Intel website, this device
claims to support DX9 and OpenGL 1.4 (wow!) - but it does vertex shaders
in the CPU (urgh!).

We didn't hold out much hope for it running WebGL - but the results were
surprising:

* In the latest build of Chrome, one of our fragment shaders (the most
complex one) causes an error in the shader linker. 
gl.getProgramParameter ( programObject, gl.LINK_STATUS) returns 'false'
- but gl.getProgramInfoLog(programObject) returns an empty string...and
gl.getError() returns gl.NO_ERROR.  Not exactly helpful!

* The latest build of Firefox/Minefield refuses to give me a graphics
context at all.

If I remove that one complex shader then my game actually runs (5Hz
full-screen!) under Chrome.  Since I'm attempting something pretty
complicated - that's actually rather impressive for a $200 computer!

It's clear that this machine is at the bottom end of what we can support
- but if Chrome were to hand me a more meaningful error code ("OMG your
shader is gigantic!") or something - then I could fall back to a less
complex shader  on that platform (normal mapping is *SO* overrated! :-)
- or at least display a more meaningful error message.

So what about Firefox?  Is it punting on context creation because the
maximum fragment shader length that the hardware can support is less
than some minimum spec or something?  Should Chrome (technically) be
punting too?

These netbooks are pretty popular - it would be nice to support them
somehow - and the one shader it's punting on is pretty outrageous.

  -- Steve

-----------------------------------------------------------
You are currently subscribed to public_webgl@khronos.org.
To unsubscribe, send an email to majordomo@khronos.org with
the following command in the body of your email: