If you have a dual core or a multiple processor machine then you can compile your programs much faster than on a single processor machine. I will show how you can compile your a Linux kernel on your Debian machine twice as fast by setting the concurrency level.

For example, in one of my previous post we saw how to measure the amount of time a particular program takes to execute. In that example, we saw it took around 10 minutes to compile a custom 2.6.28 kernel on a AMD64 Anthlon Dual core machine:

# time make-kpkg –append-to-version=-test kernel_image

real    10m58.707s
user    9m38.700s
sys    1m12.013s

However in the above example, I was not efficiently using the power of both  the processors. I simply gave the standard “make-kpkg” command to compile the kernel.

Now let’s see how we can take advantage of our dual-core machines and save us some valuable time.

# time make-kpkg –append-to-version=-test kernel_image

and here comes the best part – the results:

real    5m28.433s
user    4m18.320s
sys    1m13.276s

As we can see from the results above, we were able to compile the same kernel on the same machine in half of the time – approximately 5 minutes.

How do we find out what CONCURRENCY_LEVEL number?

Some people say that you should set the level to: Number of processors in your machine+ 1 and some say that you should set it to: Number of processors in your machine. I personally use the later.

To find out the number of processors in your machine, give the following command:

#  grep -c ^processor /proc/cpuinfo

and then you can set the concurrency level to whatever is the output of the above command. In my case it is “2″.

Now what if you are use to compile your programs/Linux kernel with the classic “make” command and you would like to make use of the concurrency level. In that case you can simple use the “-j” flag like this:

# make -j2 bzImage

Hope this will help you to save some time.

Be Sociable, Share!