RFR: 8342662: C2: Add new phase for backend-specific lowering [v6]

Quan Anh Mai qamai at openjdk.org
Wed Jan 15 02:46:42 UTC 2025


On Wed, 15 Jan 2025 02:04:58 GMT, Xiaohong Gong <xgong at openjdk.org> wrote:

>> Lowering is not idealisation so I think having backend independent lowering is fine in case we need it.
>
> Yeah, I understand it. But I think it could make this specific phase more complex.  This gives one more choice to do the arch independent IR transformation, but it's really a question what the best place should be.

This is the only place to do transformations with GVN but without Ideal so I think the distinction is clear enough to avoid potential confusion.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/21599#discussion_r1915872038


More information about the hotspot-compiler-dev mailing list