Machine specs/possible in Java

[rguiu]

Sorry if I ask for something inapropiate, I am a bit new...

What the machines where Spoj is running specifications are?

I am a bit desperate with the problem ORDERS and do not know if it is possible to solve in Java.

I generate a huge test file, 50 lines 200000 elements each( code to generate this test file bellow).

My solution works in around 15 seconds(for this file), but my computer is a 2.3Mhz. I time using the "cygwin" shell command time, not internally in the Java code.

I can get a slighly better performance with a bigger initial Heap size, I use the vm properties, -Xms64m and -Xmx64m.

Is possible to do it in Java, or it is a completely hopeless task under current Spoj conditions. The situation can be that I do things wrong, what can be solved, or that it is impossible, what is a stupid situation.

The test generator.

public class Generator {
    public static void main(String args[]) {
        System.out.println(50);
        for (int i = 0; i<50; i++) {
            System.out.println(200000);
            StringBuffer b = new StringBuffer();
            b.append("0 1 0 ");
            for (int j = 0; j < 199997; j++) {
                b.append((j-(j%3))+" ");
            }
            System.out.println(b.toString());
        }
 }

Thanks

[Adrian_Kosowski]

What the machines where Spoj is running specifications are?

Pentium II 266 MHz, 96MB RAM. The command line for running JAVA is:

/usr/local/jre/bin/java -Xms32m -Xmx32m -Xbatch -cp /usr/local/jre/lib:/solutions/tested.zip Main

and works pretty fine with most programs.

I am a bit desperate with the problem ORDERS and do not know if it is possible to solve in Java.

It is possible. I took Paweł Gawrychowski's solution (which is short and neat) and ported it to JAVA. The runtime increased from 4.66s (C++) to 9.38s (JAVA).

If you are still worried about I/O, here is a way to read a test case:

n=Integer.parseInt(in.readLine());
StringTokenizer tk = new StringTokenizer(in.readLine());
for (i=0; i<n; i++)
   input[i]=Integer.parseInt(tk.nextToken());

(in is declared at the start as BufferedReader in = new BufferedReader(new InputStreamReader(System.in)) ; )

To output the answer:

StringBuffer out=new StringBuffer();
for (i=0; i<n; i++)
{
   out.append(output[i]);
   out.append(' ');
}
System.out.println(out);

I am afraid you will have to figure out the rest for yourself

[the_one_smiley_sof]

Is there a reason the heap size is set to 32 MB? Your SPOJ Bulletin post says "you can assume you have at least 64MB of physical memory at your disposal".