DISQUS

DISQUS Hello!  The comments on this profile are unclaimed and thus are unverified.

Do they belong to you? Claim these comments.

Johan's picture

Unregistered

Feeds

aliases

  • Johan
  • Joe
  • joe
  • Alan

Johan

5 months ago

in Truce: Google and Microsoft join forces for an open white space spectrum on VentureBeat
Since when did Google and Microsoft agree on the definition of "open"? Good luck!

5 months ago

in WhiteHouse.gov takes a page from Twitter’s playbook on VentureBeat
No, he can't... he's limited to 500 characters.
1 reply
Anthony Ha's picture
Anthony Ha Exactly ... Maybe I'll send them the link to this post.

11 months ago

in Even More Incriminating Evidence in The Foxconn Debacle! on The Linuxologist
Are all of your stories ripped off from Slashdot comments?

1 year ago

in Performance Comparison - C++ / Java / Python / Ruby/ Jython / JRuby / Groovy on /var/log/mind
sorry, I'd forgotten how slow mod operator is. Even faster:

public class Chain
{
private int size;

public Chain(int size)
{
this.size = size;
}

public int kill(int nth)
{
int[] people = new int[size];
for(int i=0;i 1)
{
curCount++;
if (curCount != nth)
{
people[fillPos] = people[curPos];
fillPos++;
}
else curCount = 0;
curPos++;
if (curPos == realSize)
{
realSize -= (curPos - fillPos);
curPos = 0;
fillPos = 0;
}
}
return people[0];
}

public static void main(String[] args)
{
int ITER = 100000;
runTest(ITER);
runTest(ITER);
runTest(ITER);
runTest(ITER);
}

private static void runTest(int ITER)
{
long start = System.nanoTime();
for (int i = 0 ; i ITER ; i++)
{
Chain chain = new Chain(40);
chain.kill(3);
}
long end = System.nanoTime();
System.out.println("Time per iteration = " + ((end - start) / (ITER )) + " nanoseconds.");
}

}

1 year ago

in Performance Comparison - C++ / Java / Python / Ruby/ Jython / JRuby / Groovy on /var/log/mind
I never understood why microbenchmarks are run with bad algorithms. If performance was critical, I would not write the code the way it has been written above. Also, for Java, in performance critical situations, you should run the test multiple times to see how the JIT compilation affects the performance.

Here is the code I would write for this. It works equally well in Java, C or C++.

public class Chain
{
protected static final String versionID = "@(#) $ Id: $ ";

private int size;

public Chain(int size)
{
this.size = size;
}

public int kill(int nth)
{
int[] people = new int[size];
for(int i=0;i 1)
{
curCount++;
if (curCount % nth != 0)
{
people[fillPos] = people[curPos];
fillPos++;
}
curPos++;
if (curPos == realSize)
{
realSize -= (curPos - fillPos);
curPos = 0;
fillPos = 0;
}
}
return people[0];
}

public static void main(String[] args)
{
int ITER = 100000;
runTest(ITER);
runTest(ITER);
runTest(ITER);
runTest(ITER);
}

private static void runTest(int ITER)
{
long start = System.nanoTime();
for (int i = 0 ; i ITER ; i++)
{
Chain chain = new Chain(40);
chain.kill(3);
}
long end = System.nanoTime();
System.out.println("Time per iteration = " + ((end - start) / (ITER )) + " nanoseconds.");
}

}

1 year ago

in DreamWorks to Make Live-Action Ghost in the Shell Movie in 3D on /Film
Ah, you didnt see it? i dont get why so many people are ignoring some great anime movies/series :
Returning? Login