# GL\_APPLE\_shader \_framebuffer\_fetch

**URL:** <https://forums.imgtec.com/t/gl-apple-shader-framebuffer-fetch/1452>\
**Category:** PowerVR Insider\
**Created:** [June 21, 2012, 9:58am UTC](https://forums.imgtec.com/t/gl-apple-shader-framebuffer-fetch/1452 "2012-06-21T09:58:28Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![martin-kraus](https://avatars.discourse-cdn.com/v4/letter/m/85f322/32.png) [@martin-kraus](https://forums.imgtec.com/u/martin-kraus)\
**Post date:** [June 21, 2012, 9:58am UTC](https://forums.imgtec.com/t/gl-apple-shader-framebuffer-fetch/1452/1 "2012-06-21T09:58:28Z")

</div>

[](https://developer.apple.com/videos/wwdc/2012/?id=513)Hi, I saw at[https://developer.apple.com/videos/wwdc/2012/?id=513](https://developer.apple.com/videos/wwdc/2012/?id=513) that iOS6 devices will support a new GL extension GL\_APPLE\_shader\_framebuffer\_fetch for programmable blending. I've been thinking about this is for a bit and now I think this is a really cool extension. Thanks to Imagination for providing this feature! (Apparently the hardware supports this for a long time already.)  
  
My question is: does this extension show a similar performance as usual (non-commutative) blending on PowerVR GPUs?  
  
Thanks

---

<div class="post-metadata">

**Author:** ![JoeDavis](https://sea1.discourse-cdn.com/flex015/user_avatar/forums.imgtec.com/joedavis/32/71_2.png) [@JoeDavis](https://forums.imgtec.com/u/JoeDavis)\
**Post date:** [June 21, 2012, 1:35pm UTC](https://forums.imgtec.com/t/gl-apple-shader-framebuffer-fetch/1452/2 "2012-06-21T13:35:01Z")

</div>

Hi Martin,  
  
Both options should be comparable as they both should access the on-chip colour buffer directly, but blending operations written in your own shaders may be affected by precision conversions, how they are interpreted by the shader compiler and, of course, human error.  
  
Thanks,  
Joe