This directory contains a benchmark suite for TensorKit. Most of the benchmarks are designed to capture performance characteristics of the library, and are not intended to be used as a benchmark suite for comparing different libraries. In particular, the main goal here is to catch performance regressions and/or improvements between different versions of TensorKit.

The benchmarks are written using BenchmarkTools.jl, and the full suite can be found in the SUITE global variable defined in benchmarks.jl. Sometimes, it is useful to run only a subset of the benchmarks. To do this, you can use the --modules flag to specify which modules to run. Alternatively, you can use the TensorKitBenchmarks module directly, which is designed after BaseBenchmarks to allow for conditional loading of the benchmarks.

For a more streamlined CLI experience, you can use AirspeedVelocity.jl to run the benchmarks. The following command will run the benchmarks and compare with the current main branch:

benchpkg TensorKit \
    --rev=dirty,main \
    -o benchmark/results/ \
    -exeflags="--threads=4"

To compare with previous results, the following command can be used:

benchpkgtable TensorKit \
    --rev=dirty,main \
    -i benchmark/results/ \
    -o benchmark/results/ \