fpga_quick_ram_update VS zbasic

There are multiple ways to handle this, but I usually first covert the file to pure binary, and then convert it to hex. You could Google about how to do that, or you could use my own ruby script that does the same: https://github.com/tomverbeure/fpga_quick_ram_update/blob/main/misc/create_mif.rb

Here's sort of a generic repo you can look at which might give you some ideas.

Let me also point out that I rarely use Verilators -exe flag. I tend instead to build my designs in steps: 1) Run Verilator, 2) run make -f Vtoplvl.mk in the obj_dir directory, 3) turn my test script into an object file, and only then 4) link everything together into an executable. I've also been known for peeking at Verilator's internal variables during a simulation run--but that may be another topic entirely.

As examples: - Here's how I process libelf to provide me with an array of section pointers, each containing the address to load the values at - Here's where I then load the values into my design when I'm using an external loader via a debugging bus. - Here's another copy of the same, this time running from within Verilator rather than from within externally controlled logic.

The components found in ZBasic will split the stream into two. The key files you'll want there are dbluartsim.cpp and (again) netuart.cpp. To use dbluart, you'll need three pieces: First, you'll need to call setup() to set the baud rate, then once per clock cycle you'll want to call the operator() method--which is really just a rename for the tick() method.