Skip to main content

detect_backend

trueno_gpu::backend

Function detect_backend 

Source 
pub fn detect_backend() -> Box<dyn Backend>
Expand description

Detect best available backend

Priority order:

  1. CUDA (NVIDIA) - highest performance for NVIDIA GPUs
  2. WGPU - cross-platform fallback (Vulkan/Metal/DX12)
  3. Metal - Apple-specific (subset of WGPU)
  4. Vulkan - direct Vulkan (subset of WGPU)