Update CompressionBenchmark page sizes to realistic values (64K-1MB)

Use page sizes that reflect actual Parquet page sizes observed in practice:
64KB, 128KB, 256KB, and 1MB (the default). The 20K row-count limit
(PARQUET-1414) means most numeric columns produce pages of 78-234KB,
making the previous 8KB test point unrealistic.

Also fix JMH annotation processor path for Java 17+ compatibility
and reduce warmup/measurement iterations for faster iteration.

Performance results (master vs perf-compression-bypass branch):

Compression (ops/s, higher is better):
  Codec    | Page   | Master  | Branch  | Speedup
  SNAPPY   | 64KB   | 55,208  | 62,776  | +13.7%
  SNAPPY   | 128KB  | 28,188  | 31,692  | +12.4%
  SNAPPY   | 256KB  | 13,787  | 15,392  | +11.6%
  SNAPPY   | 1MB    |  2,493  |  2,775  | +11.3%
  ZSTD     | 64KB   |  9,127  |  9,364  |  +2.6%
  ZSTD     | 128KB  |  4,517  |  4,567  |  +1.1%
  ZSTD     | 256KB  |  2,072  |  2,161  |  +4.3%
  ZSTD     | 1MB    |    446  |    439  |  -1.6%
  LZ4_RAW  | 64KB   | 38,055  | 37,088  |  -2.5%
  LZ4_RAW  | 128KB  | 17,488  | 17,192  |  -1.7%
  LZ4_RAW  | 256KB  |  9,307  |  9,229  |  -0.8%
  LZ4_RAW  | 1MB    |  2,060  |  2,266  | +10.0%
  GZIP     | 64KB   |  1,406  |  1,457  |  +3.6%
  GZIP     | 128KB  |    643  |    652  |  +1.4%
  GZIP     | 256KB  |    316  |    321  |  +1.6%
  GZIP     | 1MB    |     77  |     78  |  +1.3%

Decompression (ops/s, higher is better):
  Codec    | Page   | Master  | Branch  | Speedup
  SNAPPY   | 64KB   | 62,553  | 68,684  |  +9.8%
  SNAPPY   | 128KB  | 31,207  | 34,199  |  +9.6%
  SNAPPY   | 256KB  | 14,737  | 16,157  |  +9.6%
  SNAPPY   | 1MB    |  3,219  |  3,581  | +11.2%
  ZSTD     | 64KB   | 35,480  | 36,241  |  +2.1%
  ZSTD     | 128KB  | 22,068  | 22,328  |  +1.2%
  ZSTD     | 256KB  | 10,910  | 10,895  |  -0.1%
  ZSTD     | 1MB    |  2,433  |  2,482  |  +2.0%
  LZ4_RAW  | 64KB   |105,142  |120,335  | +14.4%
  LZ4_RAW  | 128KB  | 52,938  | 60,533  | +14.3%
  LZ4_RAW  | 256KB  | 26,360  | 30,257  | +14.8%
  LZ4_RAW  | 1MB    |  6,155  |  7,556  | +22.8%
  GZIP     | 64KB   |  9,429  |  9,772  |  +3.6%
  GZIP     | 128KB  |  4,167  |  4,541  |  +9.0%
  GZIP     | 256KB  |  1,733  |  1,915  | +10.5%
  GZIP     | 1MB    |    405  |    452  | +11.6%

Key findings:
- SNAPPY: consistent 10-14% improvement across all page sizes
- LZ4_RAW decompression: strongest gain at 14-23% faster
- GZIP decompression: 9-12% faster at 128KB+ page sizes
- ZSTD: modest 1-4% gains (JNI library already efficient)
- Gains consistent across realistic page sizes (64K-1MB)

Author: iemejia

parquet-benchmarks/pom.xml

@@ -89,6 +89,15 @@
       <plugin>
         <groupId>org.apache.maven.plugins</groupId>
         <artifactId>maven-compiler-plugin</artifactId>
+        <configuration>
+          <annotationProcessorPaths>
+            <path>
+              <groupId>org.openjdk.jmh</groupId>
+              <artifactId>jmh-generator-annprocess</artifactId>
+              <version>${jmh.version}</version>
+            </path>
+          </annotationProcessorPaths>
+        </configuration>
       </plugin>
       <plugin>
         <groupId>org.apache.maven.plugins</groupId>

parquet-benchmarks/src/main/java/org/apache/parquet/benchmarks/CompressionBenchmark.java

@@ -54,15 +54,15 @@
 @BenchmarkMode(Mode.Throughput)
 @OutputTimeUnit(TimeUnit.SECONDS)
 @Fork(1)
-@Warmup(iterations = 3, time = 2)
-@Measurement(iterations = 5, time = 3)
+@Warmup(iterations = 2, time = 1)
+@Measurement(iterations = 3, time = 2)
 @State(Scope.Thread)
 public class CompressionBenchmark {
 
   @Param({"SNAPPY", "ZSTD", "LZ4_RAW", "GZIP"})
   public String codec;
 
-  @Param({"8192", "65536", "262144"})
+  @Param({"65536", "131072", "262144", "1048576"})
   public int pageSize;
 
   private byte[] uncompressedData;