Loop 2 GPU : Transforming Loops to OpenCL Kernels as a LLVM Pass

Lately, programmers have started to take advantage of the GPU capabilities of their systems. Still, programming for the GPU can be very hard. We are trying to hide some of this complexity from the programmer by making the compiler automatically transform embarrassingly parallel loops to GPU kernels. To this end, we have implemented a compiler pass that transforms simple loops to OpenCL kernels.