User guide# Core Concepts Piecewise constant functions Tensors Evaluation The dtype system CPU and GPU execution References Working with Tensors Creating tensors Shape and copying Type casting Joining tensors Splitting tensors Iterating Random Generation Generators Noisy trigonometric PCFs How determinism works Indexing and Masking Single-element access Slicing Assignment Full-shape masking Axis masking Creating BoolTensors Differences from NumPy Gathering Assignment with integer indices Multiple index arrays Arithmetic & Comparisons Arithmetic Comparisons Evaluation & Reductions Evaluation Reductions Combining it all Plotting A single PCF Overlaying many PCFs PCF arithmetic Highlighting the mean Persistence barcodes TDA pipeline Betti curve pipeline max_time Styling Saving and loading Saving Pickle support Loading Distances and Norms Mathematical background Distance between two PCFs Pairwise distances Cross-distances L2 kernel matrices Distance matrices Symmetric matrices Norms Using distances in machine learning Persistent Homology Background The pipeline Step 1: Point clouds Step 2: Computing persistent homology Step 3: Functional summaries Complete example The Barcode class References Point Processes Poisson point process GPU Acceleration Supported platforms Automatic CPU/GPU dispatch Controlling execution CPU and GPU execution Performance considerations