XLA
XLA is a compiler that optimizes and compiles linear-algebra and array computations for supported target hardware.
Deep Learning with JAX
A compiler and runtime for linear algebra and array programs that optimizes computations such as those produced by JAX.
- How it is used
- Accelerates JAX numerical functions by compiling composed linear-algebra computations for CPUs, GPUs, TPUs, and other accelerators.
- How deep it goes
- Explains JIT compilation, frontends, backends, operation fusion, and lowering to StableHLO and HLO.
- Examples
- Uses jnp.sum(x + y * z), jax.jit(), lowering, StableHLO, HLO, and fused GPU-kernel execution.
- What you can do after
- Apply it via JAX's jit() and pmap(), then inspect lowered representations before compiling numerical functions.
Deep Learning with R, Third Edition
XLA is a compiler for accelerating linear algebra and machine-learning computations across supported hardware.
- How it is used
- Accelerates TensorFlow functions after their behavior has been debugged in the R runtime.
- How deep it goes
- Explains the compilation switch and startup cost, but stops short of compiler internals.
- Examples
- Uses a ReLU dense function containing matrix multiplication and bias addition.
- What you can do after
- Apply it via TensorFlow’s tf_function() by setting jit_compile = TRUE after debugging.
Deep Learning with Python, Third Edition
Accelerated Linear Algebra, TensorFlow's optional compiler used to further optimize and speed up function execution.
- How it is used
- Optimizes TensorFlow function execution when faster repeated calls justify additional compilation time.
- How deep it goes
- Explains activation, first-call overhead, and later speedups, but does not descend into compiler internals.
- Examples
- Demonstrates XLA by compiling a TensorFlow dense function containing matrix multiplication and ReLU.
- What you can do after
- Apply it via TensorFlow's jit_compile=True option after debugging the function in Python.
Read how Deep Learning with Python, Third Edition explains XLA →
Related concepts
- JAXused by · JAX is a Python library for high-performance array-based numerical computing with composable transformations including automatic differentiation, compilation, vectorization, and parallelization.
- just-in-time compilationused by · Just-in-time compilation translates or optimizes program code during execution, often using runtime information to generate specialized executable instructions.
- parallel mapused by · A parallel map is an operation that applies a computation concurrently across inputs or processing devices and returns the corresponding results.
- Jaxprused by · JAX's simple intermediate representation of a traced function, used as the input to later transformations and compilation.
- kernel fusionaddresses · Kernel fusion is an optimization that combines multiple computational operations into fewer executable units to reduce launch overhead and intermediate data movement.
- TPUuses · A Tensor Processing Unit (TPU) is a specialized processor designed to accelerate tensor computations, particularly for machine-learning workloads.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.