cuTENSOR: A High-Performance CUDA Library For Tensor Primitives

Welcome to the cuTENSOR library documentation.

cuTENSOR is a high-performance CUDA library for tensor primitives; its key features are:

• Extensive mixed-precision support:

  • FP64 inputs with FP32 compute.

  • FP32 inputs with FP16, BF16, or TF32 compute.

  • Complex-times-real operations.

  • Conjugate (without transpose) support.

• Support for up to 12-dimensional tensors.

• Arbitrary data layouts.

• Trivially serializable data structures.

• Main computational routines:

  • Direct (i.e., transpose-free) tensor contractions.

  • Tensor reductions (including partial reductions).

  • Element-wise tensor operations:

    • Support for various activation functions.

    • Arbitrary tensor permutations.

    • Conversion between different data types.

The documentation consists of three main components:

• A User Guide that introduces important basics of cuTENSOR including details on notation and accuracy.

• A Getting Started guide that steps through a simple tensor contraction example.

• An API Reference that provides a comprehensive overview of all library routines, constants, and data types.

Contents:

• User Guide
  • Nomenclature
  • Einstein Notation
  • Performance Guidelines
  • Software-managed Plan Cache (beta)
  • Accuracy Guarantees
  • Scalar Types
  • Supported GPUs
  • Environment Variables
• Getting Started
  • Installation and Compilation
  • Headers and Data Types
  • Define Tensor Sizes
  • Initialize Tensor Data
  • Create Tensor Descriptors
  • Create Contraction Descriptor
  • Determine Algorithm and Workspace
  • Plan and Execute
• Plan Cache (beta)
  • Incremental Autotuning
  • Introductory Example
  • Advanced Example
• API Reference
  • cuTENSOR Data Types
  • cuTENSOR Functions

Indices And Tables

• Index

• Search Page