This new implementation provides a constant time computation regardless of the length of B, rather than the previous implementation which took linear time relative to the length of B.

#### Benchmark summary

CPU: AMD Ryzen Threadripper PRO 3945WX 12-Cores
OS: Debian Testing
```
# JMH version: 1.37
# VM version: JDK 11.0.17, OpenJDK 64-Bit Server VM, 11.0.17+8
# VM invoker: /home/benwhitehead/opt/java/jdk-11.0.17+8/bin/java
# VM options: -Xms1g -Xmx1g -XX:+AlwaysPreTouch
# Blackhole mode: full + dont-inline hint (auto-detected, use -Djmh.blackhole.autoDetect=false to disable)
# Warmup: 4 iterations, 30 s each
# Measurement: 8 iterations, 30 s each
# Timeout: 10 min per iteration
# Threads: 1 thread, will synchronize iterations
# Benchmark mode: Sampling time
```

```
Benchmark                                  (objectSize)    Mode      Cnt     Score    Error  Units
Crc32cBenchmark.concatCrc32c                     131072  sample  9243752     0.074 ±  0.001  us/op
Crc32cBenchmark.concatCrc32c                     262144  sample  9015106     0.078 ±  0.001  us/op
Crc32cBenchmark.concatCrc32c                     524288  sample  9010257     0.074 ±  0.001  us/op
Crc32cBenchmark.concatCrc32c                    1048576  sample  9165519     0.074 ±  0.001  us/op
Crc32cBenchmark.concatCrc32c                    2097152  sample  8993038     0.075 ±  0.001  us/op
Crc32cBenchmark.crc32cCombineGoogle              131072  sample  2070639   115.806 ±  0.018  us/op
Crc32cBenchmark.crc32cCombineGoogle              262144  sample  1029209   233.019 ±  0.045  us/op
Crc32cBenchmark.crc32cCombineGoogle              524288  sample   516517   464.360 ±  0.106  us/op
Crc32cBenchmark.crc32cCombineGoogle             1048576  sample   262118   915.067 ±  0.354  us/op
Crc32cBenchmark.crc32cCombineGoogle             2097152  sample   131385  1825.685 ±  0.834  us/op
```