I’m a Julia starter and I would want to leverage the language in doing cyber-physical simulation in the embedded systems domain. Something similar to what Renode can do but dynamically (if we can think of how to run the application software source instead of its compiled binary ELF or HEX), but also supports running these binaries.
Please let me know if someone is already doing the same and suggestions on how to do it.
There is work on compilation for embedded simulations, hardware-in-the-loop, etc. that is being developed as part of JuliaSim - Julia Computing, though more details are not public as of now. If you represent a interested party and are interested in being a commercial partner please feel free to contact me privately.
Is the compilation part going to be closed source? IE generating a standalone small c binary or something like that (similar to tf lite).
Useful for autonomous vehicles or edge deployment, not Sciml style simulations
Thanks. What about other more standard static compilation stories? (ie facilities for tree shaking and minimal runtime to deploy webservers and CLI apps with smaller memory footprints and lower latency than is currently possible.)
There are some things that can be improved in the Julia compiler (and are already being done), and some things which need a completely alternative compiler pipeline. But getting to full embedded can require runtime free scenarios.
Isn’t that basically GPUCompiler.jl?