It’s faster on a GPU

Sometimes they are right but if you haven’t tried it on an FPGA or don’t know much about FPGAs then how can you be Wholesale NFL Jerseys right or have any My credibility.  If someone tells you that an algorithm is better on an GPU ask them how much they know about FPGAs.  Not to challenge them but I cheap nfl jerseys think cheap nfl jerseys most cases I have come across are just regurgitation of GPU marketing!  Remember they market heavily and FPGA companies nfl jerseys china don’t.  There are cases Ray Ban sale where the FPGAs will be better so let’s leave the FUD out of this and let the numbers speak for themselves.

2 Responses to “It’s faster on a GPU”

  1. John says:

    Well I’ve done the math and even the most ideal FPGA implementation will be about two orders of magnitude slower than an equivalently priced set of of GPUs. The issue lies in the density of FPGAs and how pricey they are because of that. Also at low volumes the FPGA makers are unable to use the latest process technologies that are available to the likes of nVidia and ATI, inflating the size of their silicon even more.

    I wish you the best of luck though. It’s definitely a good project to have in your portfolio if you can come up with a working implementation.

    Source: Computer Engineering student, having implemented both a CPU and a GPU on an FPGA.

    • terry says:

      Would be good to get more info and numbers on your math; is it published anywhere? There are too many factors to oversimplify things the way you do.

Leave a Reply

Your email address will not be published. Required fields are marked *