jextract: bytes passing benchmarks

[ktoso]

On an M2 macbook, I'll run later on linux as well but this is just about rough ideas:

Benchmark                                      (totalBytes)  Mode  Cnt       Score       Error  Units
JNIByteArrayBenchmark.acceptBuffer_jni                 4096  avgt    3       0.314 ±     0.214  us/op
JNIByteArrayBenchmark.acceptBuffer_jni                65536  avgt    3       1.085 ±     0.112  us/op
JNIByteArrayBenchmark.acceptBuffer_jni             16777216  avgt    3     287.848 ±    73.858  us/op
JNIByteArrayBenchmark.acceptBytes_jni                  4096  avgt    3      42.275 ±   115.404  us/op
JNIByteArrayBenchmark.acceptBytes_jni                 65536  avgt    3     628.038 ±   215.706  us/op
JNIByteArrayBenchmark.acceptBytes_jni              16777216  avgt    3  159460.417 ± 60075.214  us/op
JNIByteArrayBenchmark.acceptData_jni                   4096  avgt    3       0.015 ±     0.001  us/op
JNIByteArrayBenchmark.acceptData_jni                  65536  avgt    3       0.016 ±     0.005  us/op
JNIByteArrayBenchmark.acceptData_jni               16777216  avgt    3       0.015 ±     0.001  us/op
JNIByteArrayBenchmark.acceptMutableBuffer_jni          4096  avgt    3       0.340 ±     0.042  us/op
JNIByteArrayBenchmark.acceptMutableBuffer_jni         65536  avgt    3       3.109 ±     0.312  us/op
JNIByteArrayBenchmark.acceptMutableBuffer_jni      16777216  avgt    3     616.923 ±   467.034  us/op
JNIByteArrayBenchmark.acceptNested_jni                 4096  avgt    3      41.110 ±     7.999  us/op
JNIByteArrayBenchmark.acceptNested_jni                65536  avgt    3     629.788 ±   201.350  us/op
JNIByteArrayBenchmark.acceptNested_jni             16777216  avgt    3  156828.319 ± 16632.867  us/op
JNIByteArrayBenchmark.denseShard_jni                   4096  avgt    3      43.136 ±     3.939  us/op
JNIByteArrayBenchmark.denseShard_jni                  65536  avgt    3     630.803 ±   333.042  us/op
JNIByteArrayBenchmark.denseShard_jni               16777216  avgt    3  160196.253 ± 18796.422  us/op
JNIByteArrayBenchmark.echoBytes_jni                    4096  avgt    3      40.404 ±     5.682  us/op
JNIByteArrayBenchmark.echoBytes_jni                   65536  avgt    3     635.765 ±    37.211  us/op
JNIByteArrayBenchmark.echoBytes_jni                16777216  avgt    3  162764.906 ± 42880.991  us/op
JNIByteArrayBenchmark.echoData_jni                     4096  avgt    3       0.191 ±     0.102  us/op
JNIByteArrayBenchmark.echoData_jni                    65536  avgt    3       0.303 ±     4.236  us/op
JNIByteArrayBenchmark.echoData_jni                 16777216  avgt    3       0.189 ±     0.160  us/op
JNIByteArrayBenchmark.echoNested_jni                   4096  avgt    3      42.729 ±    10.072  us/op
JNIByteArrayBenchmark.echoNested_jni                  65536  avgt    3     633.071 ±   156.590  us/op
JNIByteArrayBenchmark.echoNested_jni               16777216  avgt    3  159919.885 ± 24323.902  us/op
JNIByteArrayBenchmark.returnBytes_jni                  4096  avgt    3       0.560 ±     0.307  us/op
JNIByteArrayBenchmark.returnBytes_jni                 65536  avgt    3       2.797 ±     7.527  us/op
JNIByteArrayBenchmark.returnBytes_jni              16777216  avgt    3     631.423 ±   354.456  us/op
JNIByteArrayBenchmark.returnData_jni                   4096  avgt    3       1.019 ±     0.231  us/op
JNIByteArrayBenchmark.returnData_jni                  65536  avgt    3      10.607 ±    11.934  us/op
JNIByteArrayBenchmark.returnData_jni               16777216  avgt    3    2516.995 ±  2224.608  us/op
JNIByteArrayBenchmark.returnNested_jni                 4096  avgt    3       2.423 ±     1.649  us/op
JNIByteArrayBenchmark.returnNested_jni                65536  avgt    3       5.400 ±     7.372  us/op
JNIByteArrayBenchmark.returnNested_jni             16777216  avgt    3     604.660 ±    65.900  us/op
JNIByteArrayBenchmark.sparseShard_jni                  4096  avgt    3     728.087 ±   188.162  us/op
JNIByteArrayBenchmark.sparseShard_jni                 65536  avgt    3     738.260 ±    99.180  us/op
JNIByteArrayBenchmark.sparseShard_jni              16777216  avgt    3     728.695 ±   209.163  us/op