LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Programming (https://www.linuxquestions.org/questions/programming-9/)
-   -   information about performance on cipher algorithms? (https://www.linuxquestions.org/questions/programming-9/information-about-performance-on-cipher-algorithms-870693/)

xeon123 03-24-2011 07:00 AM

information about performance on cipher algorithms?
 
Hi,

I would like to know how does it take a cipher algorithm to compute a message. So, I'm looking for information that shows benchmark results on cipher algorithms. Does anyone can point me a link?

rtmistler 03-24-2011 08:31 AM

Insert time stamps into your code and make logs before, during, and after the cipher computations.

Also, find a number of viable test cases. Repeating data, random data, short data, long data, patterned data, to determine if the algorithm works better or worse depending on what gets fed into it.

If you have the code for the algorithm, you can predict the number of instructions that get executed, determine the exact number of clock cycles required for them all, and more precisely determine the exact CPU time required.

Tinkster 03-24-2011 11:35 AM

Moved: This thread is more suitable in <PROGRAMMING> and has been moved accordingly to help your thread/question get the exposure it deserves.

xeon123 03-24-2011 11:43 AM

Inserting timestamps is a good idea, but I'm looking for official benchmark cipher algorithms tests.

rtmistler 03-24-2011 11:51 AM

I'm sure there are some benchmark programs out there for certain things; however I'll bet that you'll have difficulty finding exactly what you're looking for, hence why I suggested building your own benchmark test.

If you can find the author of the cipher algorithms you do have, contact them and see if they made any performance test utilities along with their program. For that matter, they may have included that in their source already. Search their code for any references to tests or test modes.

If you're evaluating several binary packages, then I would design my own tests so that I could uniformly and fairly evaluate the performance of each of the algorithms so that I could get all the information I'd require. Plus if I had written the tests, then they would be portable from system to system and I could compare Windows to Linux or different CPU and memory combinations on various machine types.


All times are GMT -5. The time now is 08:00 AM.