Error compiling vecadd example

[lmzuccarelli]

First of all thanks for rebooting this project.

I followed the prerequisites but used cuda 12.9 , installed llvm 7.1.0 set the LLVM_CONFIG envar and ensured libnvvm was also set.

I'm getting the following error (undefined symbol)

error: failed to run custom build command for `vecadd v0.1.0 (/home/lzuccarelli/Projects/rust-kernel-cuda)`

Caused by:
  process didn't exit successfully: `/home/lzuccarelli/Projects/rust-kernel-cuda/target/release/build/vecadd-c3ca7fb11c812cab/build-script-build` (exit status: 101)
  --- stdout
  cargo::rerun-if-changed=build.rs
  cargo::rerun-if-changed=kernels
  cargo:rerun-if-changed=/home/lzuccarelli/Projects/rust-kernel-cuda/kernels

  --- stderr
  error: failed to run `rustc` to learn about target-specific information

  Caused by:
    process didn't exit successfully: `/home/lzuccarelli/.rustup/toolchains/nightly-2025-03-02-x86_64-unknown-linux-gnu/bin/rustc - --crate-name ___ --print=file-names -Zcodegen-backend=/home/lzuccarelli/Projects/rust-kernel-cuda/target/release/deps/librustc_codegen_nvvm.so '-Zcrate-attr=feature(register_tool)' '-Zcrate-attr=register_tool(nvvm_internal)' -Zcrate-attr=no_std -Zsaturating_float_casts=false '-Cllvm-args=-arch=compute_61 --override-libm' --target nvptx64-nvidia-cuda --crate-type bin --crate-type rlib --crate-type dylib --crate-type cdylib --crate-type staticlib --crate-type proc-macro --print=sysroot --print=split-debuginfo --print=crate-name --print=cfg -Wwarnings` (exit status: 1)
    --- stderr
    error: couldn't load codegen backend /home/lzuccarelli/Projects/rust-kernel-cuda/target/release/deps/librustc_codegen_nvvm.so: undefined symbol: _ZN4llvm6Triple9normalizeB5cxx11ENS_9StringRefE


  thread 'main' panicked at /home/lzuccarelli/Projects/Rust-CUDA/crates/cuda_builder/src/lib.rs:542:10:
  Did not find output file in rustc output

Could you perhaps give me some pointers ? I looked into the cuda_builder crate but to be honest I could not really get to the root of the problem

Thanks in advance

[LegNeato]

Hm, weird! Does docker work for you?

[lmzuccarelli]

TBH I haven't tried it. I'll give that a go.

[LegNeato]

Any luck with https://github.com/Rust-GPU/Rust-CUDA/pkgs/container/rust-cuda-ubuntu24-cuda12?

[lmzuccarelli]

No luck - still navigating problems with podman (I use the nvidia container toolkit's nvidia-ctk cdi generate) to ensure podman picks up GPU's but its not detecting my GPU, the joys of not using docker :)

[LegNeato]

I can't repro on any of my vast.ai machines, either when installing llvm7.1 directly or when using the container. Please feel free to reopen or file again!