Jump to content

  • Log In with Google      Sign In   
  • Create Account

Banner advertising on our site currently available from just $5!

1. Learn about the promo. 2. Sign up for GDNet+. 3. Set up your advert!


Member Since 09 Sep 2005
Offline Last Active Nov 12 2014 08:39 AM

Posts I've Made

In Topic: OpenCL OpenGL interop

29 April 2013 - 07:09 AM

1) clSetKernelArg(kernel, 0, sizeof(vbo), &vbo) should be: clSetKernelArg(kernel, 0, sizeof(cl_mem), &PositionArray) 2) remove clWaitForEvents(1, &wait) (it is not necessary) 3) clEnqueueNDRangeKernel can take NULL as a local size argument

In Topic: Parallel Octree Construction on GPU

30 September 2012 - 03:08 AM

You usually convert the BVH construction problem to a sorting problem. This can be done with the use of morton codes, see for example this paper http://mgarland.org/...pers/gpubvh.pdf Bitonic sort and Radix sorts can be implemented very efficiently on modern GPUs,
see here for example for an OpenCL implementation: http://www.bealto.co...ting_intro.html

In Topic: Tessellation makes point

19 June 2012 - 08:52 AM

Do you set the primitives to GL_PATCHES and set the glPatchParameteri ?

In Topic: Need help with reliable polygon triangulation

03 November 2010 - 12:15 AM

Another vote for the Triangle program. In my experience it can easily handle cases like 2 and 3. Although the interface is a little cumbersome (especially if you try to specify internal holes), it is very robust and fast. To AvengerDr: yes the result can be index based.

In Topic: Water Ripples

23 September 2010 - 01:54 AM

I am not familiar with the language you use, but I do not think that the casting from a float to a byte will work. You must find the maximum,minimun height and scale the values or do some clamping. Anyway, some blokiness is expected to that kind of approximation algorithms. Personally, I use a better approximation based on this paper.