Print tune parameters with `--dlaf:print-config`

[msimberg]

From the --dlaf:print-config perspective, the tune parameters and config options are the same. Should we print them together (in one block) or separate them (with a heading?).

[msimberg]

cscs-ci run

[msimberg]

cscs-ci run

[msimberg]

LGTM!

Do we need a comment in include/dlaf/tune.h to remember to update operator<< when a new parameter is added?

That's a good idea. See https://github.com/eth-cscs/DLA-Future/pull/990/commits/9e479c0e4a35cca77ea9c66611330e5cd5709aad. Does that look like a useful comment to you? I also added the same comment for configuration options.

[msimberg]

cscs-ci run

[msimberg]

cscs-ci run

[msimberg]

This does not yet contain #989 so if all goes well this PR should now fail in CI with the assertions added in https://github.com/eth-cscs/DLA-Future/pull/990/commits/ddfbf0a5bfee299e4cee18408e0fe3c8db1d00c1. Will rebase once (if) it fails.

[msimberg]

This does not yet contain #989 so if all goes well this PR should now fail in CI with the assertions added in ddfbf0a. Will rebase once (if) it fails.

https://gitlab.com/cscs-ci/ci-testing/webhook-ci/mirrors/4700071344751697/7514005670787789/-/jobs/5204882236#L1886

59: [ERROR] /DLA-Future/src/init.cpp:158 : void dlaf::internal::updateConfigurationValue(const pika::program_options::variables_map&, T&, const string&, const string&) [with T = long int; std::string = std::__cxx11::basic_string] 59: env_var.find("DLAF") == std::string::npos 59: DLAF_BT_BAND_TO_TRIDIAG_HH_APPLY_GROUP_SIZE 59: 59: terminate called without an active exception

[msimberg]

cscs-ci run

[msimberg]

cscs-ci run