Skip to content
Benchmarking and performance tracking for adcc
Python Shell Vim script
Branch: master
Clone or download

Latest commit

Fetching latest commit…
Cannot retrieve the latest commit at this time.

Files

Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
benchmarks
scripts
templates
.gitignore
LICENSE
README.md
asv.conf.json
setup.cfg
setup_environment.sh

README.md

Benchmarking and performance tracking tools for adcc

This repository is a basic benchmarking suite for adcc based on airspeed velocity.

Running the benchmarks

This illustrates how to run the benchmarks against the current master of adcc. For running against a different commit, see the details of airspeed velocity.

First install airspeed velocity and ccache (e.g. apt install ccache or brew install ccache). Then activate the ccache environment and run the benchmarks:

. setup_environment.sh
asv run

If you are running on a cluster, where you can spare some computational time, you can set the environment variable ADCC_BENCH_EXPENSIVE, e.g.

export ADCC_BENCH_EXPENSIVE=yes

to enable running the expensive benchmarks as well.

Scope of the benchmark suite

Currently the benchmark suite covers a few individual steps of the ADC calculation as well as multiple fulrun testcases, where both SCF and ADC is run on a problem. For the latter cases both computational time and memory usage is benchmarked.

The fullrun ADC test cases include:

  • Small atoms and molecules: Water, silane, neon
  • Medium-sized molecules: Paranitroaniline, Noradrenaline

Adding new fullrun testcases

Adding new fullrun benchmark cases can be achieved by adding a molecule to the file benchmarks/Cases.py.

You can’t perform that action at this time.