I have written a compiler for a language at more or less this abstraction level. It provides access to 16 general purpose registers and a set of virtual instructions that operate on these registers. I program on an intel mac so the virtual instructions all map directly to x86_64 instructions, but it would be very straightforward to write an arm backend that may compose multiple arm instructions for the equivalent x86_64 behavior. I also could support virtual registers for platforms with fewer than 16 registers.

Having this compiler, which is extremely fast (it is self hosted and self compiles its own 9750K line source file in 15ms on my 2019 macbook pro and I've seen throughout ranging from 500K to 10M lines per second for other programs), I have little interest in ever using LLVM again. I would rather just write optimal code than pray that LLVM does a good job for me. For equivalent programs, I find LLVM can be anywhere from 50-10000x slower than my compiler and doesn't necessarily produce better code. I can produce examples of LLVM code becoming significantly worse at -O2 than at -O0.

The only real utility LLVM offers me at this point is that I could use it to figure out what an optimal solution to a small subprogram might be by compiling a small c program and inspecting the output. Then I can just directly write the optimal solution instead of being perpetually bogged down by llvm's bloat as well as also being exposed to unexpected regressions in the optimizer and/or having to learn its strange incantations to get desired performance.

> Something in between C and non-portable assembly.

Writing your own LLVM passes?