I am working on a design to do the real time functions that typically done in HDL in a processor optimized to run C in as few cycles as possible. I really need input whether it sounds desirable. Initial reactions in other groups hasn't gotten past the laughing stage, because everyone already knows it can't be done.
I have an incomplete but compilable bdf that shows the hardware resources are small and the memory used of course will be a function of code size.
The kind of configuration I visualize is an Nios running the application with several small processors controlling the peripherals and managing data flow. The question is how useful is it? The same thing can be done with Nios, but I think too expensive and this design will take fewer cycles.
Any code samples would be appreciated. I don't have C fully implemented so samples would help to see if a subset is acceptable. For instance implementing a call stack for recursive functions will be slower than inline functions.
Here is a link where you can download more info:
http://mysite.verizon.net/vzeosqt4/embeddedcengine/ Much thanks for any help.
