RFR: 8332670: C1 clone intrinsic needs memory barriers

[Aleksey Shipilev]

On Wed, 5 Jun 2024 15:12:33 GMT, Galder Zamarreño <galder at openjdk.org> wrote:

> Let me make sure I understand this right: you are suggesting removing the `if (zero_array)` branch and instead always emit storestore barrier, as well as the storestore barrier at the end of `append_alloc_array_copy`?

Yes. AFAICS, we would have an excess `StoreStore` only on C1 clone path, which is should be fast enough already after you introduced the intrinsic. That is to say, that even if that `StoreStore` regresses performance a little bit in comparison to current mainline tip, it should still be faster than the version that did not do C1 clone intrinsic implemented at all.

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

PR Comment: https://git.openjdk.org/jdk/pull/19538#issuecomment-2150348432