The challenge is that the Vision C5 DSP needs to be incorporated into an SoC. Testing can be done on simulators but off-the-self platforms are not currently available whereas there are many GPGPU ...

However, GPGPU boards can process far more parallel data using thousands of CUDA cores. So, the power-to-performance ratio is what’s affected in a very positive way.

XtremeDSP Starter Kit – Spartan-3A DSP FPGA Edition. (Click this image to view a larger, more detailed version) Also included is System Generator for DSP, which enables the use of The Mathworks ...

If you're gonna do CUDA only, the 600 series isn't bad, but for OpenCL, you'd be better off with Fermi (500 series and to an extent, I wouldn't pitch a GT520 against a GTX680 in GPGPU and expect ...