[code-reflection] RFR: Experiment: Alternative SSA construction algorithm

[Paul Sandoz]

On Wed, 14 Aug 2024 07:58:53 GMT, Hannes Greule <hgreule at openjdk.org> wrote:

> Hi,
> 
> this is mainly an experiment, I wanted to play around with the API and see how an alternative SSA construction algorithm can be implemented. There are a few notable problems I encountered:
> 
> - The original algorithm is supposed to work on-the-fly in a single step. This does not work with the current API.
>   - `Value`s (operands) and `Reference`s (successor arguments) are bound eagerly. I assume terminating ops could be emitted only after everything else was done, but that's not easy to achieve currently either. For operands, I don't see any way to say "oh, please use a different value" after the operation is passed to the builder. I'm not sure how relevant this is for other transformations, but if we find a simple solution to this without giving up immutability, I could imagine it to be very powerful.
>   - I'm currently using more data structures to cope with the two-step variant than the original algorithm itself requires.
>     - The `predecessors` map could be avoided at the cost of additional computation whenever only filled blocks should be considered
>     - The `loads` map is required as we can't replace the loads directly during the analysis step
>     - The `additionalParameters` could probably be removed if `Block.Builder` allowed removing parameters
>     - The `deletedPhis` set is also needed to cope with the fact that we look up current defs in both steps
> - I noticed that for the method in `TestTraverse`, this implementation gets rid of one parameter that the existing implementation does not get rid of. I'm not sure if it is designed to produce minimal SSA, but that looks odd. (`c` is passed to the while body, but it is never read there before being written)
> 
> 
> I think the algorithm itself is pretty straightforward, but the extra steps needed to adapt to the API somewhat nullify that.
> I tried to document the implementation details that differ from the paper, but if there are any questions or suggestions how this could be improved, please let me know.

(BTW the changes from `Set` to `SequencedSet` are nice, feel free to separate that out as a separate PR.)

(Still working my way through the paper + code, i'l send some further thoughts later today.)

Explaining phis at that level is i would imagine a proxy for explaining control flow join points. If one thinks of blocks as little functions, that jump to other functions as a tail call, it might actually be easier to explain.

> Combined with the ability to remove block parameters again, it would potentially made some things a little easier.

This should be possible to support, the `Block.Builder::parameters` currently returns an unmodifiable list. I was hesitant about exposing a mutable list, since removing a parameter that is used will create an invalid model. But, we could support removing block parameters that are not currently used nor are any successor arguments associated with them. Although, that probably means it is not very useful.

In general my approach to the design as been to lean into multiple passes (analysis/transformation) rather than try and optimize for single passes. That's a trade off. We are not trying to build a framework for compiler pipelines like say LLVM, MLIR, or C2, although there are obvious similarities to key aspects.

-------------

PR Comment: https://git.openjdk.org/babylon/pull/214#issuecomment-2293804522