Using integrated GPU and EXLA?

matt-savvy · August 31, 2025, 7:53pm

Hey gang, I’m running a Livebook locally on a machine that has an onboard graphics card.

Processor: AMD Ryzen™ 7 PRO 6850U with Radeon™ Graphics × 16

I’m seeing some mixed answers about if this is supported by the rocm backend or not.

I can try to set the default backend client like so:

Nx.default_backend({EXLA.Backend, client: :rocm})

But the error message I’m seeing seems like it’s looking for ”cuda”

** (exit) exited in: GenServer.call(EXLA.Client, {:client, :rocm, [platform: :rocm]}, :infinity)
    ** (EXIT) an exception was raised:
        ** (RuntimeError) Could not find registered platform with name: "cuda". Available platform names are: Host

Anyone run into this before? Am I just missing some config?

polvalente · September 1, 2025, 3:02pm

github.com/elixir-nx/xla

Support for ROCM 6

opened 05:06PM - 13 May 24 UTC

jalberto

It seems ROCM 5.6 kind of works, but it really requires too much back and forth… to have everything working, the new Fedora 40 brings official ROCM support but starting in ROCM 6. I am using this config from https://github.com/elixir-nx/xla/issues/63 ```elixir Mix.install( [ {:web_driver_client, "~> 0.2.0"}, {:kino, "~> 0.12.3"}, {:req, "~> 0.4.14"}, {:erlexec, "~> 2.0"}, {:nx, github: "elixir-nx/nx", sparse: "nx", override: true}, {:exla, github: "elixir-nx/nx", sparse: "exla", override: true} ], system_env: %{ "XLA_ARCHIVE_URL" => "https://static.jonatanklosko.com/builds/0.6.0/xla_extension-x86_64-linux-gnu-rocm.tar.gz", "ROCM_PATH" => "/usr/lib64/rocm/" }, config: [nx: [default_backend: {EXLA.Backend, client: :host}]] ``` I managed to find every pkgs it was asking for (this took a while of back and forth) until I reached this: ```elixir 18:36:37.767 [warning] The on_load function for module Elixir.EXLA.NIF returned: {:error, {:load_failed, ~c"Failed to load NIF library /home/ja/.cache/mix/installs/elixir-1.16.2-erts-14.2.5/f3927a87654a1bf097d7e31b6277a9f8/_build/dev/lib/exla/priv/libexla: 'librocblas.so.3: cannot open shared object file: No such file or directory'"}} ``` My guess is `xla_extension` needs to be built for rocm 7 (librocblas.s0.4), I tried to build it myself but the requirements are too way off the current system (gcc versions and so on) Will be great if there were official xla binaries for different ROCM versions, as there are for CUDA. I understand ROCM support is in low priority, but it is really nice for start in AI as it works nicely in linux

This is a relevant issue to this question