Irwin D'Souza

AOT changes look good to me. @hzongaro do you mind being committer for this one?

@jdmpapin could you also review the JIT changes from an optimizer pov?

@LinHu2016 sorry, in the formatting changes I requested I meant that your code was using tabs, whereas the compiler code uses spaces. As it stands, the formatting is still incorrect...

FYI @a7ehuo @jdmpapin feel free to edit the desc if there's anything wrong / I've missed anything.

> Considering that the offsets are sensitive to whether or not anything earlier is flattened, it might be better to identify fields by something like an index Right, the idea...

> From your setup here with u and n, the break-even point is u = n/2. Your original scheme will be smaller when u < n/2, and my scheme will...

fyi @mpirvu @vijaysun-omr @mstoodle

There are some non-trivial technical difficulties associated with continuing the process but throwing away the JIT'd code. For one, we would need to run in a mode similar to (but...

I recall in OCP, for example, you could choose what nodes you wanted deployments to run on. If only some of the machines are too old, then you could just...

> Why would that portable code use any features that might be absent on CPUs that the JIT normally supports? Well, AFAICT the [support statement](https://www.eclipse.org/openj9/docs/openj9_support/) doesn't speak about CPUs. As...