error[E0463]: can't find crate for `core` `nvptx64-nvidia-cuda` target may not be installed

[symbioteGod]

Trying to create an Ubuntu/CUDA Docker image. First, I installed the CUDA/Docker toolkit. Then, I created these 2 attempts at Dockerfiles:

https://github.com/symbioteGod/tor-v3-vanity_Docker/blob/master/Dockerfile

https://github.com/symbioteGod/tor-v3-vanity_Docker/blob/master/Dockerfile_bak

but I always get the error:

   Compiling ed25519-dalek v1.0.1
error: failed to compile `tor-v3-vanity v0.1.0 (/root/tor-v3-vanity)`, intermediate artifacts can be found at `/root/tor-v3-vanity/target`

Caused by:
  failed to run custom build command for `tor-v3-vanity v0.1.0 (/root/tor-v3-vanity)`

Caused by:
  process didn't exit successfully: `/root/tor-v3-vanity/target/release/build/tor-v3-vanity-f763f96eafbe3b4e/build-script-build` (exit code: 1)
  --- stdout
  cargo:rustc-link-search=native=/usr/local/cuda/lib64/

  --- stderr
  [PTX] Unable to build a PTX crate!
  [PTX] warning: unused manifest key: build
  [PTX]  Downloading crates ...
  [PTX]   Downloaded quote v1.0.7
  [PTX]   Downloaded proc-macro2 v1.0.19
  [PTX]   Downloaded byteorder v1.3.4
  [PTX]   Downloaded ed25519-compact v0.1.6
  [PTX]   Downloaded syn v1.0.36
  [PTX]    Compiling proc-macro2 v1.0.19
  [PTX]    Compiling unicode-xid v0.2.1
  [PTX]    Compiling syn v1.0.36
  [PTX]    Compiling byteorder v1.3.4
  [PTX]    Compiling rustacuda_core v0.1.2
  [PTX]    Compiling ed25519-compact v0.1.6
  [PTX] error[E0463]: can't find crate for `core`
  [PTX]   |
  [PTX]   = note: the `nvptx64-nvidia-cuda` target may not be installed
  [PTX] 
  [PTX] error: aborting due to previous error
  [PTX] 
  [PTX] For more information about this error, try `rustc --explain E0463`.
  [PTX] error: could not compile `rustacuda_core`
  [PTX] 
  [PTX] warning: build failed, waiting for other jobs to finish...
  [PTX] error[E0463]: can't find crate for `core`
  [PTX]   |
  [PTX]   = note: the `nvptx64-nvidia-cuda` target may not be installed
  [PTX] 
  [PTX] error: aborting due to previous error
  [PTX] 
  [PTX] For more information about this error, try `rustc --explain E0463`.
  [PTX] error: build failed
The command '/bin/sh -c rustup toolchain add nightly-2020-01-02 && rustup target add nvptx64-nvidia-cuda --toolchain nightly-2020-01-02;cd ${INSTALL_PATH}/tor-v3-vanity;rustup install nightly;rustup target add nvptx64-nvidia-cuda;cargo install ptx-linker;cargo +nightly install --path .' returned a non-zero code: 101
The terminal process "/usr/bin/zsh '-c', 'docker build --pull --rm -f "Dockerfile" -t tor-v3-vanity-cuda:latest "."'" terminated with exit code: 101.

[JKamsker]

Same on windows

[satyr-software]

Try : rustup +nightly target add nvptx64-nvidia-cuda cargo +nightly install ptx-linker

[VigilantBag]

I am going to take a guess and say we are using windows as our host. Do NOT install the deb file. Install using the runfile according to the nvidia forums https://forums.developer.nvidia.com/t/failure-to-install-cuda-on-wsl-2-ubuntu/128592/8. The deb file installs the native linux driver which you do not want on WSL. At the same time, you need to be running an insider preview build of windows (>20211) in order for the windows kernel driver to communicate with WSL2.

[TheFiZi]

Try : rustup +nightly target add nvptx64-nvidia-cuda cargo +nightly install ptx-linker

This solved my problem on my Ubuntu 20.04.3 LTS VM (not WSL)