Emil J

Actually, this repo seems abandoned. Development seems to have moved to a fork at https://github.com/NatronGitHub/Natron which is where my package manager must have gotten it from. Even your 2.4.2 can't...

Interesting. Here's my reproducer: [repro.tar.gz](https://github.com/riscv-software-src/riscv-isa-sim/files/13239211/repro.tar.gz) Let `cc` be clang or gcc of your choice `cc scream.c htif.c a.s -nostdlib -Tvanilla-spike.ld -march=rv32g` `spike --isa=RV32G a.out > /dev/null` will not surpress output

Can't reproduce, adding `std::cerr

If you set RISCV_GCC_OPTS in your make invocation by removing that flag from the default value of RISCV_GCC_OPTS in benchmarks/Makefile, this error should go away. It might break the tests...

I can't address this immediately but I hope to play with riscv-tests some more next week and let you know

I have followed this advice and it worked for spike and qemu but it looks like a very bad idea. I encountered a bug when I adapted it for building...

> I'm not really sure what you are trying to do. As a dev for an embedded toolchain vendor, I would like to provide simple "hello world" support code to...

I have found a more proper solution in [riscv-coremark](https://github.com/riscv-boom/riscv-coremark/blob/master/riscv64-baremetal/syscalls.c). The lowRISC code was missing a pair of `__sync_synchronize();` barriers. The noinline attribute is no longer necessary

Hard to say what could be wrong without seeing your command line invocations of compiler and spike. My guess would be you didn't use matching -march -mabi etc flags for...

The README provides instructions and a simple test: `spike pk hello`. If this doesn't work for you, maybe open a new, specific issue. If you don't want to use pk,...