Follow up on noise data for study group

[ ] Are machines quiet (no unnecessary services, etc..)
[X] Categorize noisy benchmarks by memory
[ ] How do we deal with tests that change from noisy to not noisy over time
[ ] Verify confidence in large oscialltion tests
[X] Compare runs across different machines to see if we see impact across them
[ ] How do we use the data to reduce noise (Build our test suite using these to have a set of noisy tests)
[ ] Possibly look into measuring windows defender in some way during test run
[ ] Cut inter machine data across test to check for slow machines
[ ] Use linux queue for inter-machine noise

dotnet / performance