• silenium_dev@feddit.org
    link
    fedilink
    arrow-up
    1
    ·
    1 month ago

    If you’re doing anything with GPU compute (Blender, AI, simulations etc.), just ROCm, CUDA or oneAPI alone will take up half of that