Support manually provided benchmark results

[dclong]

Some benchmarks might require priviledged permission and cannot be run in GitHub Actions VMs. In those situations, accepting manually provided benchmark results (run elsewhere) can be useful.

[miparnisari]

Isn't this already supported? https://github.com/benchmark-action/github-action-benchmark#:~:text=Additionally%2C%20even%20though%20there%20is%20no%20explicit%20example%20for%20them%2C%20you%20can%20use%20customBiggerIsBetter%20and%20customSmallerIsBetter%20to%20use%20this%20action%20and%20create%20your%20own%20graphs%20from%20your%20own%20benchmark%20data.%20The%20name%20in%20these%20tools%20define%20which%20direction%20%22is%20better%22%20for%20your%20benchmarks.