FPGAs offer the promise of reconfigurable computing: Reconfiguring the hardware to match your application needs. This reconfigurable approach is often more efficient than general purpose processors such as CPUs. However, programming FPGAs has traditionally been hard and usually required expertise in Hardware Description Languages (HDLs). Altera has introduced an OpenCL SDK for FPGAs, making them accessible to a much larger software community and a potential alternative to other accelerators such as GPGPUs. We take a look at Altera's offering, explain how it works and provide analysis about how using OpenCL for FPGAs compares with alternatives such as GPUs.
via CG Class @ NCSU http://cgclass.posthaven.com/find-an-interesting-idea-opencl-for-programming-fpgas