RFR: 8353741: Improve UUID.toString performance by using SIMD within a register instead of table lookup

[Chen Liang]

On Mon, 6 Jan 2025 13:18:50 GMT, Shaojin Wen <swen at openjdk.org> wrote:

> Improve the performance of UUID::toString by using Long.expand and SWAR (SIMD within a register) instead of table lookup. Eliminating the table lookup can also avoid the performance degradation problem when the cache misses.

Does this expandNibbles compile to any xor nodes in the GVN?

Note that aarch64 has an intrinsic for expand that goes through the vector processing unit. Don't know if that intrinsic or the constant folding takes priority; might need further tweaks to hotspot if that intrinsic comes before the constant folded result and slows things down.

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

PR Comment: https://git.openjdk.org/jdk/pull/22928#issuecomment-2589380162