I ran across both a Tweet covering an article on NextPlaform covering Amazon's usage of FPGAs in F1 and a free eBook which is available for a week. The title of the eBook is FPGA Frontiers: New Applications and Trends in Reconfigurable Computing and it is only available for free until Jan 27, 2017. So go grab your copy now -- yes I know that since this eBook was sponsored by Xilinx it has some propaganda inside, but the discussion is worthwhile in my humble opinion.
As I referenced in my post about our R&D actions with Maxeler Technologies I think there is a broader market trend going on here: more usage of various chip types. This is leading me to think that perhaps we need some serious R&D around compilers, development environments, workload managers, memory access models, and even the humble operating system scheduler. These kinds of developments don't happen over night and are in fact quite intense with input from academics, startups and the open source community. However, given that compilers and the lot aren't really sexy at this time I believe we'll see not so much VC funding or grants heading this way for a bit.
What do you think? Do you believe we need advances in these core technologies or are we ok? If so are there any companies already doing this for even just the 4 types of chips that Intel has to offer?