MINOR: Optimize bit-packing decoder performance by increasing unpackCount in Packer16

[1111nit]

Rationale for this change The current Packer16 decoder implementation using JDK Vector API has performance limitations with unpackCount = 16 for bit-packing decoding.

What changes are included in this PR? Modified ByteBitPacking512VectorLE.java to increase the processing batch size: // Before private static final int UNPACK_COUNT = 16; // After private static final int UNPACK_COUNT = 32;

Testing Details Testing environment:

• JDK: OpenJDK 17.0.2
• Test framework: JMH Benchmark
• Platform: x86_64

Performance results:

• Baseline throughput: 29.587 ops/s
• Optimized throughput: 54.524 ops/s
• Performance gain: 84.3% improvement

[wgtmac]

cc @jatin-bhateja