ee_u32 accessed with %d format specifier about coremark

[petertorelli]

Good to see 16-bit getting some representation! I think option 2, PRIu32 is the best solution since the intent is to print a 32-bit value. I'll queue this up for the next release, we don't make changes very often, so for the short term I recommend just modifying the printf. I have encountered a case of someone running CoreMark for huge numbers of iterations during testing, so >64Ksec isn't unheard of.

What compiler and target are you using, out of curiosity?

[mysterymath]

https://github.com/llvm-mos/llvm-mos

LLVM backend for the MOS 6502!

It's still really early days with respect to optimization work (so don't post this anywhere too official ;) ), but I'm happy to report that a 1MHz MOS 6502 achieves 0.0386 CoreMarks/s with our backend.

Looking at the generated output, there's a ton of obvious inoptimalities. It's nice to have such a comprehensive benchmark to work against! Porting it was surprisingly easy, too.