CODEWITHSUNDEEP
javaperformancefilemd5checksum
lucky_start_izumi
lucky_start_izumi

Reputation: 2591

Very slow to generate MD5 for large file using Java

I am using Java to generate the MD5 hash for some files. I need to generate one MD5 for several files with a total size of about 1 gigabyte. Here's my code:

private String generateMD5(SequenceInputStream inputStream){
    if(inputStream==null){
        return null;
    }
    MessageDigest md;
    try {
        int read =0;
        byte[] buf = new byte[2048];
        md = MessageDigest.getInstance("MD5");
        while((read = inputStream.read(buf))>0){
            md.update(buf,0,read);
        }
        byte[] hashValue = md.digest();
        return new String(hashValue);
    } catch (NoSuchAlgorithmException e) {
        return null;
    } catch (IOException e) {
        return null;
    }finally{
        try {
            if(inputStream!=null)inputStream.close();
        } catch (IOException e) {
            // ...
        }
    }

}

This seems to run forever. How can I make it more efficient?

Upvotes: 10

Views: 22528

Answers (3)

kriegaex
kriegaex

Reputation: 67297

Whenever speed is an issue and you download a file from a URL and want to calculate its MD5 at the same time (i.e. not save the file, reopen and read again just to get its MD5), my solution at https://stackoverflow.com/a/11189634/1082681 might be helpful. It is based on Bloodwulf's code snippet here in this thread (thanks!) and just extends it a bit.

Upvotes: 0

Alanmars
Alanmars

Reputation: 1277

I rewrite your code with nio, the code is somewhat like below:

private static String generateMD5(FileInputStream inputStream){
    if(inputStream==null){

        return null;
    }
    MessageDigest md;
    try {
        md = MessageDigest.getInstance("MD5");
        FileChannel channel = inputStream.getChannel();
        ByteBuffer buff = ByteBuffer.allocate(2048);
        while(channel.read(buff) != -1)
        {
            buff.flip();
            md.update(buff);
            buff.clear();
        }
        byte[] hashValue = md.digest();
        return new String(hashValue);
    }
    catch (NoSuchAlgorithmException e)
    {
        return null;
    } 
    catch (IOException e) 
    {
        return null;
    }
    finally
    {
        try {
            if(inputStream!=null)inputStream.close();
        } catch (IOException e) {

        }
    } 
}

On my machine, it takes about 30s to generate md5 code for a large file, and of course i test your code as well, the result indicates that nio doesn't improve the performance of the program.

Then, i try to get the time for io and md5 respectively, the statistics indicates that the slow file io is the bottleneck because about 5/6 of time is taken for io.

By using the Fast MD5 library mentioned by @Sticky, it takes only 15s to generate md5 code, the improvement is remarkable.

Upvotes: 12

Sticky
Sticky

Reputation: 1082

You may want to use the Fast MD5 library. It's much faster than Java's built-in MD5 provider and getting a hash is as simple as:

String hash = MD5.asHex(MD5.getHash(new File(filename)));

Be aware that the slow speed may also be due to slow File I/O.

Upvotes: 22

Related Questions