actuarial/basicterm_me_python repository overview

This image is built in the CI of the https://github.com/actuarialopensource/benchmarks⁠ repository.

It runs the BasicTerm_ME calculations from lifelib. By default, it runs 10,000 modelpoints. This dockerfile allows you to repeat these modelpoints to test over larger modelpoint sets. If --multiplier 10000 then there will be 10,000 times the base modelpoints, or 100,000,000 modelpoints.

To run this on GPU:

docker run --gpus all actuarial/basicterm_me_python:latest --multiplier 10000 --model jax_iterative
docker run --gpus all actuarial/basicterm_me_python:latest --multiplier 10000 --model torch_recursive

On CPU omit the --gpus all. Probably don't try to run 100 million modelpoints on CPU it will be slow.

docker run actuarial/basicterm_me_python:latest --multiplier 100 --model jax_iterative