I am familiar with C++ and OpenGL 1.2.
Platform: Windows64 / Linux64
I want to give a try to OpenCL for my matrix library (multiplication, addition, gauss-jordan elimination etc). I work with big matrices (e.g. 50000x50000).
These are some starter questions:
[ul]
[li]Ok, headers provided. What about library? Where is opencl.a? For OpenGL, libraries are provided with MingW. But for OpenCL?[/3mt2o44o][/li][li]If opencl.a provided from hardware vendor, how can we be sure that the program will work on other vendors?[/
3mt2o44o][/li][li]A OpenCL program run only on GPUs or run on CPUs too?[/
3mt2o44o][/li][li]If I am a freak and I have 2 different GPUs in one machine, it will work with both of them “correctly” or I must wait for the slowest GPU?[/
3mt2o44o][/li][]How big is the floating point? 32bit? 64bit? 128bit?[/
3mt2o44o][/ul]