Here's one where I was testing the timing behaviour of a hubRAM write access when colliding with the FIFO's hubRAM accesses. I've specified that the assembly be truly inlined as hubexec code along with the compiled code. Side effect - it doesn't incur the block copy to cogRAM overhead when truly inlined.
PS: The "FIFO" acts like an instruction prefetcher when a program is executing from hubRAM, AKA hubexec.
Comments
Here's an example I did recently for testing:
Here's one that loads a spin file and uses the methods from it.
Here's one where I was testing the timing behaviour of a hubRAM write access when colliding with the FIFO's hubRAM accesses. I've specified that the assembly be truly inlined as hubexec code along with the compiled code. Side effect - it doesn't incur the block copy to cogRAM overhead when truly inlined.
PS: The "FIFO" acts like an instruction prefetcher when a program is executing from hubRAM, AKA hubexec.