954,518 Members — Technology Publication meets Social Media
Username:
Password:
Lost login information?
Have something to say? Contribute New Article Reply to this Article

how do you make java programs faster?

This is just a general question. I was wondering how you can make programs compile and run quicker? Does it have to do with the primative type you use or what?

server_crash
Postaholic
2,111 posts since Jun 2004
Reputation Points: 113
Solved Threads: 20
 

ibm.com has an article about how to make java programs faster ... I dont remember the page address ... but you can search ibm for java.

nanosani
Unauthenticated Liar
Team Colleague
1,830 posts since Jul 2004
Reputation Points: 45
Solved Threads: 56
 

hello everyone,
I once did this with the awt class in which i developed a small program in C++ and an exact clone of it in Java (functionality wise). I also developed another program in C++ to act as a benchmark thus testing seeing how fast each program responds to anything.

All the java functions were done in different threads including its click events

It seems that the Java is faster than C++ ranging from 330ms-390ms for all functionalities.

This was done years ago.........

I don't even know where the C++ benchmark program or it source codes are anymore
I am not sure about swing as i only tested for awt

Yours Sincerely

Richard West

freesoft_2000
Practically a Master Poster
623 posts since Jun 2004
Reputation Points: 25
Solved Threads: 10
 

http://www-106.ibm.com/developerworks/java/library/j-javaopt/#1


I found the link for performance improvement.

nanosani
Unauthenticated Liar
Team Colleague
1,830 posts since Jul 2004
Reputation Points: 45
Solved Threads: 56
 

Really the best way to speed up your program's compile time is simply getting a faster compute machine. I don't really see the point in inproving the compile time as you are not taking hours to compile are you?

For the run time improvements, you should have a good grasp on how java works. This mostly comes through common sense, a solid computer science background, and experience.

There are many tools you could use: profilers, memory leak checkers, etc. If you are developing for the enterprise, Wily is a good tool I've used before: http://www.wilytech.com/

Ah and also, a good site I like to look at for reference on Java performance tuning is of course: http://www.javaperformancetuning.com/


Ed

cosi
Junior Poster
153 posts since Aug 2004
Reputation Points: 17
Solved Threads: 1
 

freesoft: It seems highly likely that your test was flawed. spawning new threads for every gui action is generally a bad idea.

There are many factors that could have caused the performance you saw. You say you wrote the same functionality in the two programs---that doesn't say much. A sloppily written Java program is just as bad as a sloppily written C++ program. It doesn't sound like you created the same program in C++ and java.

That said, I believe that for most programs written with care in C++ and Java, you should expect roughly the same performance.


Ed

hello everyone, I once did this with the awt class in which i developed a small program in C++ and an exact clone of it in Java (functionality wise). I also developed another program in C++ to act as a benchmark thus testing seeing how fast each program responds to anything.

All the java functions were done in different threads including its click events

It seems that the Java is faster than C++ ranging from 330ms-390ms for all functionalities.

This was done years ago.........

I don't even know where the C++ benchmark program or it source codes are anymore I am not sure about swing as i only tested for awt

Yours Sincerely

Richard West

cosi
Junior Poster
153 posts since Aug 2004
Reputation Points: 17
Solved Threads: 1
 

This article has been dead for over three months

Post: Markdown Syntax: Formatting Help
You