Skip to content

Unable to detect GPU #76

@RTMorris1

Description

@RTMorris1

Hello,
I am using a Linux machine running Ubuntu 20.04 with a NVIDIA GeForce RTX 4060 Ti.
the PATH includes /usr/local/cuda-12.8/bin
the LD_LIBRARY_PATH includes /usr/local/cuda-12.8/lib64

When I run nvcc --version I get the following:
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2025 NVIDIA Corporation
Built on Fri_Feb_21_20:23:50_PST_2025
Cuda compilation tools, release 12.8, V12.8.93
Build cuda_12.8.r12.8/compiler.35583870_0

When I installed RcppML and ran the gpu_available() command the result is FALSE.
When I load the torch library and ran the cuda_device_count() command the result is 1, so it can detect the GPU.
Why doesn't RcppML detect an available GPU?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions