Reduce allocations

[shawjef3]

This reduces allocations due to not boxing every Long in ZScheduler#submittedLocations. I measured the difference by running ZIO.unit.forever and profiling the same way I did at #7365.

benchmarks

before

[info] Benchmark                                            Mode  Cnt     Score    Error  Units
[info] ZSchedulerBenchmarks.catsRuntimeChainedFork         thrpt   15  6207.667 ± 65.372  ops/s
[info] ZSchedulerBenchmarks.catsRuntimeForkMany            thrpt   15   670.941 ±  9.549  ops/s
[info] ZSchedulerBenchmarks.catsRuntimePingPong            thrpt   15   986.860 ± 31.994  ops/s
[info] ZSchedulerBenchmarks.catsRuntimeYieldMany           thrpt   15   855.601 ± 18.382  ops/s
[info] ZSchedulerBenchmarks.zioFixedThreadPoolChainedFork  thrpt   15  2735.637 ± 34.485  ops/s
[info] ZSchedulerBenchmarks.zioFixedThreadPoolForkMany     thrpt   15   389.630 ±  3.819  ops/s
[info] ZSchedulerBenchmarks.zioFixedThreadPoolPingPong     thrpt   15   784.109 ±  6.717  ops/s
[info] ZSchedulerBenchmarks.zioFixedThreadPoolYieldMany    thrpt   15    14.408 ±  0.143  ops/s
[info] ZSchedulerBenchmarks.zioSchedulerChainedFork        thrpt   15  2609.220 ±  8.991  ops/s
[info] ZSchedulerBenchmarks.zioSchedulerForkMany           thrpt   15   444.072 ±  3.386  ops/s
[info] ZSchedulerBenchmarks.zioSchedulerPingPong           thrpt   15  1886.767 ±  8.340  ops/s
[info] ZSchedulerBenchmarks.zioSchedulerYieldMany          thrpt   15    67.884 ±  0.189  ops/s

after

[info] Benchmark                                            Mode  Cnt     Score    Error  Units
[info] ZSchedulerBenchmarks.catsRuntimeChainedFork         thrpt   15  6266.270 ±  5.935  ops/s
[info] ZSchedulerBenchmarks.catsRuntimeForkMany            thrpt   15   664.790 ±  1.493  ops/s
[info] ZSchedulerBenchmarks.catsRuntimePingPong            thrpt   15   936.911 ±  9.402  ops/s
[info] ZSchedulerBenchmarks.catsRuntimeYieldMany           thrpt   15   725.193 ±  2.423  ops/s
[info] ZSchedulerBenchmarks.zioFixedThreadPoolChainedFork  thrpt   15  2806.162 ±  6.056  ops/s
[info] ZSchedulerBenchmarks.zioFixedThreadPoolForkMany     thrpt   15   417.146 ±  1.950  ops/s
[info] ZSchedulerBenchmarks.zioFixedThreadPoolPingPong     thrpt   15   715.587 ±  3.365  ops/s
[info] ZSchedulerBenchmarks.zioFixedThreadPoolYieldMany    thrpt   15    14.125 ±  0.080  ops/s
[info] ZSchedulerBenchmarks.zioSchedulerChainedFork        thrpt   15  2917.393 ±  6.393  ops/s
[info] ZSchedulerBenchmarks.zioSchedulerForkMany           thrpt   15   445.130 ±  6.354  ops/s
[info] ZSchedulerBenchmarks.zioSchedulerPingPong           thrpt   15  1863.503 ± 13.140  ops/s
[info] ZSchedulerBenchmarks.zioSchedulerYieldMany          thrpt   15    71.960 ±  0.133  ops/s

allocations

before

after

Notice that there are many fewer java.lang.Long allocations.

[adamgfraser]

Thank you! 🙏

[shawjef3]

@adamgfraser you're welcome. It says it's ready to merge, but I'm unable to click the button myself.