Intel MKL

Q: How deeply does Fast Python: High performance techniques for large datasets cover Intel MKL?

Explains backend selection, threading consequences, configuration inspection, and installation, but stops short of benchmark results.

As explained in Fast Python: High performance techniques for large datasets

Intel MKL, now Intel oneMKL, is a library of optimized mathematical and scientific computing routines, including implementations of BLAS and LAPACK.

Intel MKL, now Intel oneMKL, supplies optimized BLAS and LAPACK routines for numerical and scientific computing. NumPy can use it beneath array operations and linear-algebra algorithms, including matrix multiplication and singular value decomposition. section 4.3.1

The linked BLAS/LAPACK implementation changes how NumPy uses CPU resources. MKL is threaded, so parallel workloads require care to avoid overcommitting CPU resources; a nonthreaded library instead leaves process-level parallelism more predictable. section 4.3.1

Use configuration inspection before tuning:

Run np.show_config() to inspect the libraries and directories associated with the NumPy installation. section 4.3.1
When configuration output is unclear, inspect package files and symbolic links; libcblas.so.3 pointing to libmkl_rt.so indicates MKL linkage. section 4.3.1
Avoid the slow, nonthreaded NetLib implementation when performance matters; compare OpenBLAS and MKL for maximum system-specific performance. section 4.3.2

conda create -n book-mkl blas=*=mk
conda activate book-mkl
conda install numpy

Creates an isolated Anaconda environment with the MKL BLAS build and installs NumPy without changing the existing environment. section 4.3.2

Checking the active backend

Run np.show_config(), then verify the reported library paths rather than inferring the backend from an environment name. If necessary, inspect the installed shared-library links directly. section 4.3.1

How Fast Python: High performance techniques for large datasets covers it

Cover of Fast Python: High performance techniques for large datasets

liveBook manning.com

Read Fast Python: High performance techniques for large datasets on liveBook Get the book at manning.com

Related concepts

CPUused by · A central processing unit (CPU) is a general-purpose processor that executes computer instructions and performs computational operations.
BLASconfigures · BLAS is a standard interface and collection of low-level routines for efficient vector and matrix operations in numerical computing.
OpenBLASalternative to · OpenBLAS is an open-source optimized implementation of the BLAS and LAPACK numerical linear algebra interfaces.

This explainer was generated by Manning's knowledge graph from the text of Fast Python: High performance techniques for large datasets. The section links point to the parts of the book it draws on.