Java 内存映射文件MappedByteBuffer

MappedByteBuffer是java nio引入的文件内存映射方案,读写性能极高。在NIO中主要用到普通的输入流,带缓冲的输入流,RandomAccessFile和MappedByteBuffer。

现在我们来看看这四种流的效率,废话少说直接上代码。

我们采用CRC32来循环冗余校验。CRC32在java.util.zip包中。

1.普通的输入流。

public static long checksumInputStream(Path filename) throws IOException
  {
    try (InputStream in = Files.newInputStream(filename))
    {
        CRC32 crc = new CRC32();
 
        int c;
        while ((c = in.read()) != -1)
          crc.update(c);
        return crc.getValue();
    }
  }

2.带缓冲的输入流

public static long checksumBufferedInputStream(Path filename) throws IOException
{
  try (InputStream in = new BufferedInputStream(Files.newInputStream(filename)))
  {
      CRC32 crc = new CRC32();
 
      int c;
      while ((c = in.read()) != -1)
        crc.update(c);
      return crc.getValue();
  }
}

3.RandomAccessFile

public static long checksumRandomAccessFile(Path filename) throws IOException
  {
    try (RandomAccessFile file = new RandomAccessFile(filename.toFile(), "r"))
    {
        long length = file.length();
        CRC32 crc = new CRC32();
 
        for (long p = 0; p < length; p++)
        {
          file.seek(p);
          int c = file.readByte();
          crc.update(c);
        }
        return crc.getValue();
    }
  }

4.MappedByteBuffer

public static long checksumMappedFile(Path filename) throws IOException
{
  try (FileChannel channel = FileChannel.open(filename))
  {
      CRC32 crc = new CRC32();
      int length = (int) channel.size();
      MappedByteBuffer buffer = channel.map(FileChannel.MapMode.READ_ONLY, 0, length);
 
      for (int p = 0; p < length; p++)
      {
        int c = buffer.get(p);
        crc.update(c);
      }
      return crc.getValue();
  }
}

每个流对应的方法写好之后我们来开始测试

public static void main(String[] args) throws IOException {
 
        String name = "txt.txt";
        Path filename = Paths.get(name);
        long start, crcValue, end;
 
        System.out.println("Input Stream:");
        start = System.currentTimeMillis();
 
        crcValue = checksumInputStream(filename);
        end = System.currentTimeMillis();
        System.out.println(Long.toHexString(crcValue));
        System.out.println((end - start) + " milliseconds");
 
        System.out.println("Buffered Input Stream:");
        start = System.currentTimeMillis();
        crcValue = checksumBufferedInputStream(filename);
        end = System.currentTimeMillis();
        System.out.println(Long.toHexString(crcValue));
        System.out.println((end - start) + " milliseconds");
 
        System.out.println("Random Access File:");
        start = System.currentTimeMillis();
        crcValue = checksumRandomAccessFile(filename);
        end = System.currentTimeMillis();
        System.out.println(Long.toHexString(crcValue));
        System.out.println((end - start) + " milliseconds");
 
        System.out.println("Mapped File:");
        start = System.currentTimeMillis();
        crcValue = checksumMappedFile(filename);
        end = System.currentTimeMillis();
        System.out.println(Long.toHexString(crcValue));
        System.out.println((end - start) + " milliseconds");
    }

为了保证每种流相互之间对结果没有影响,我们可以分别把main方法中其他流的代码注释掉。

最后可以看到MappedByteBuffer效率最高,所消耗的时间最少。

猜你喜欢

转载自www.linuxidc.com/Linux/2015-09/123533.htm