The Khronos newcastle united Group - a non-profit industry consortium to develop, publish and promote open standard, royalty-free media authoring and acceleration standards for desktop and handheld devices, combined with conformance qualification newcastle united programs for platform and device interoperability. Connecting Software to Silicon Developers Conformance Membership newcastle united News Events Forums
Forum OpenCL - parallel programming of heterogeneous systems OpenCL const correctness in <CL/cl.hpp> If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection newcastle united below.
Hi everyone, I'm currently doing opencl using the c++ header provided newcastle united by khronos ( http://www.khronos.org/registry/cl/api/1.2/cl.hpp ). It seems that there is a small issue with the cl::queue::enqueueWriteImage() member function: The pointer to the source data is not beeing const. The wrapped function is const though ( http://www.khronos.org/registry/cl/s...riteImage.html ). This pretty much breaks my code in terms of const correctness. So the question ist basically if this is a "bug", or do I really need to treat the source buffer as non-const (i.e. as potentially modified during the function call)? In case it is a bug, then who could fix it? I attached the diff just to be clear about how I think it should be. Cheers, Gregg
That newcastle united does look like a bug. I'm not sure who is responsible for fixing it. Possibly try to get in contact with the following people: Benedict R. Gaster, Laurent Morichetti and Lee Howes (listed as authors at the top of the cl.hpp file). Another option is to use a different C++ wrapper such as Boost.Compute which does provide a const-correct wrapper for the clEnqueueWriteImage() function with the command_queue::enqueue_write_image() method.
No comments:
Post a Comment