Add tests for expected results

[timcassell]

To more easily validate changes to toolchains/auto generated code, we want to be able to run some tests to make sure we get expected results for all runtimes and toolchains. Empty method should have 0 (or less than 1 cpu cycle) time, overhead should not take more time than workload.

CI machines are unreliable due to the nature of their hardware and shared resources, so these will need to be part of the manual tests rather than part of the normal tests.

We can also validate memory allocations since it will be manual tested (we had memory measurement issues failing tests in the past).