Phyx
This would be unfortunate as it would mean we would effectively stop tracking performance on e.g. Windows and Mac OS since the current implementation doesn't allow for the data to live together in the same repo.
The implementation as it stands does indeed allow data from multiple platforms to live side-by-side. Each measurement from a given testsuite run carries with it a "test environment" which is meant to allow us to distinguish between various operating system/architecture/hardware configurations. This defaults to "local" but can be overridden on the `make test` command line. The CI builders will do this to provide with unique, descriptive names, giving us a running history of the test characteristics on all of our CI platforms. Cheers, - Ben