Code and data for benchmark of GTApprox
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
figures
testcases
README.md
plotResults.py
runAllTests.py
runSingleTest.py

README.md

GTApprox benchmark

These are scripts and data we used to compare GTApprox against several popular approximation libraries, as described in the paper GTApprox: surrogate modeling for industrial design (journal publication, arXiv preprint).

RRMS error profile Training time profile

Run runAllTests.py to save results for all individual tests. (Each test is run in a separate process by first restricting available virtual memory (with ulimit, Linux-only) and then calling the script runSingleTest.py.) After completion, run plotResults.py to collect results and plot performance profiles.