Array copy optimization

Question

0.00/5 (No votes)

See more:

Good day guys

I'm inserting an array into another array with the Array.copy function. The destination array is fixed. Its a byte[157]. The first 64 elements and the last 77 elements stay the same only the middle 16 elements change. So I say:

<br />
Array.copy(source,0,dest,64,16);<br />

Isn't there a faster way of doing this? I know about memoryStreams but the memoryStream.toArray() eats up performance and it creates a copy of the array which uses memory I don't have. I'm developing on a smart card. Maybe the full code will help. This is a Hmac function used for tls.

C#

private void ComputeMyHMac_hash(ref HashAlgorithm alg, ref byte[] secret, byte[] seed, ref byte[] output,int size)
        {
            #region pads
            #region inpad
            byte[] inpad = { 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36,
                             0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36,
                             0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36,
                             0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36,
                             0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36,
                             0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36,
                             0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36,
                             0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 
                           };
            #endregion
            #region outpad
            byte[] outpad = { 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C,
                             0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C,
                             0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C,
                             0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C,
                             0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C,
                             0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C,
                             0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C,
                             0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 0x5C, 
                           };
            #endregion
            for (int i = 0; i < secret.Length; i++)
            {
                inpad[i] ^= secret[i];
                outpad[i] ^= secret[i];
            }
            #endregion
            //end init pads
            secret = null;
            byte[] buf = new byte[size];
            byte[] buf2 = new byte[size];
            byte[] worker = new byte[64+size];
            Array.Copy(inpad, worker, inpad.Length);
            byte[] worker2 = new byte[64  + seed.Length];
            Array.Copy(inpad,worker2,inpad.Length);
            Array.Copy(seed, 0, worker2, 64, seed.Length);
            //var for doFinal
            byte[] tmp = new byte[size];
            byte[] worker3 = new byte[64 + size];
            Array.Copy(outpad, worker3, outpad.Length);
            outpad = null;
            
            MemoryStream outStr = new MemoryStream(0);
            int iterations = (output.Length + size - 1) / size;
            for (int i = 0; i < iterations; i++)
            {
                if (i == 0)
                {
                    DoFinal(ref buf, worker2, ref alg, ref tmp, ref worker3);
                    worker2 = new byte[64 + size + seed.Length];
                    Array.Copy(inpad, worker2, inpad.Length);
                    inpad = null;
                    Array.Copy(seed, 0, worker2, 64 +size, seed.Length);
                    ////debug.WriteLine("ComputeMyHMac_hash 1");
                }
                else
                {
                    Array.Copy(seed, 0, worker, 64, seed.Length);
                    DoFinal(ref buf, worker, ref alg, ref tmp, ref worker3);
                }
                seed = buf;
                Array.Copy(seed, 0, worker2, 64, seed.Length);
                DoFinal(ref buf2, worker2, ref alg, ref tmp, ref worker3);
                outStr.Write(buf2, 0, buf2.Length);
            }
            output = outStr.ToArray();
            outStr.Close();
        }
private void DoFinal(ref byte[] buf, byte[] worker, ref HashAlgorithm alg, ref byte[] tmp, ref byte[] worker3)
        {
            tmp = alg.ComputeHash(worker);
            Array.Copy(tmp, 0, worker3, 64, tmp.Length);
            buf = alg.ComputeHash(worker3);
        }

The hash algorithm can be MD5 or Sha1, The size is the size of the hash algorithm either 16 or 20.

Hopefully someone can help. Any advice would be appreciated.
Thanks

Btw: Where can I find the editing tags to make the post look good? To put the code in a nice code block that the indentation is correct?

Posted 14-Oct-10 21:06pm

Derberos

Updated 14-Oct-10 21:28pm

Dalek Dave

v2

Add a Solution

Comments

Dalek Dave 15-Oct-10 3:28am

Edited for Readability

Henry Minute 15-Oct-10 3:46am

Highlight existing code then click on 'code block' at the top of the editor. If you paste code in it should get blocked automatically, if not check the state of the radio buttons at the bottom.

1 solution

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

E.F. Nijboer · Answer 1 · 2010-10-14T21:45:00

You could make your own copy function that uses a larger data type. If you can be sure that the start and number of items to copy are both always dividable by four, you can use UInt32 for example. If it's always dividable by eight you can use UInt64 (using MMX registers). And if you are actually sure it's always dividable by sixteen you could also use an 128bit data type but you then you would have to write your own assembly code to utilize the XMM registers (and before use, you must make sure they are available.)

There is however a trade of in using the MMX and XMM registers in small operations. You would be forced to use the EMMS instruction and on an earlier pentium 4 this is a very excessive cycle consuming instruction.

In conclusion it would be best to use UInt32 because it doesn't involve any assembly and other trade-offs there might be. As long as the index and the number of bytes to copy are divisible by four there won't be any problem but would speed up copying 4 times because it's moving 4 bytes instead of 1 at a time.

Good luck!