I am newbie about GPGPU and OpenCL. I wonder if I can use OpenCL to speedup an elaboration (signal processing) on a Texas Instrument omap3530 which is equipped with PoverVR SGX 5 serie GPU (it should be 520). Omap3530 is a cortex A8 CPU running Linux, and we will write our algorithms in C/C++.

Note that omap3530 also integrate a dsp but we don't want to use it.