Inline stage 1

I'm manually inlining all vector operations. This is the result of a discussion on [Panama's mailing list](https://mail.openjdk.org/pipermail/panama-dev/2024-June/020445.html). ### 256-bit vectors | Benchmark | main (ops/s) | inline_stage1 (ops/s) | |--------------------------------------------------------------------------------|--------------|-----------------------| | ParseAndSelectBenchmark.countUniqueUsersWithDefaultProfile | 955.428 | 990.783 | | ParseAndSelectBenchmark.countUniqueUsersWithDefaultProfile (padded) | 983.696 | 1048.394 | | SchemaBasedParseAndSelectBenchmark.countUniqueUsersWithDefaultProfile | 1436.906 | 1566.293 | | SchemaBasedParseAndSelectBenchmark.countUniqueUsersWithDefaultProfile (padded) | 1536.843 | 1707.806 | ### 512-bit vectors | Benchmark | main (ops/s) | inline_stage1 (ops/s) | |--------------------------------------------------------------------------------|--------------|-----------------------| | ParseAndSelectBenchmark.countUniqueUsersWithDefaultProfile | 1736.325 | 1838.197 | | ParseAndSelectBenchmark.countUniqueUsersWithDefaultProfile (padded) | 1851.136 | 1931.859 | | SchemaBasedParseAndSelectBenchmark.countUniqueUsersWithDefaultProfile | 2973.474 | 3146.485 | | SchemaBasedParseAndSelectBenchmark.countUniqueUsersWithDefaultProfile (padded) | 3163.667 | 3512.966 |