pub fn pack_vertices_from_xy(
device: &Arc<Device>,
queue: &Arc<Queue>,
inputs: &Scatter2GpuInputs,
params: &Scatter2GpuParams,
) -> Result<GpuVertexBuffer, String>Expand description
Builds a GPU-resident vertex buffer for scatter plots directly from provider-owned XY arrays. Only single-precision tensors are supported today; callers should fall back to host packing for other types.