Ccycler a CPU benchmark

[sulairris]

so i made a little one click CPU benchmark in java.

 

ready to run

 

running

 

done

 

 

you can get it here for free http://nathanlp.tumblr.com/Ccycler
if you use it post your score.

[Lolzious]

uh.. okay?

How does it work?

[the jolly roger]

self advertising is against the CoC

[Tedster]

Okay, I'll virus scan it then try it...

[sulairris]

uh.. okay?

How does it work?

It does a crap ton of math and times how long it takes to do it.

[Guest]

It does a crap ton of math and times how long it takes to do it.

 

What kind of math.

[Lolzious]

It does a crap ton of math and times how long it takes to do it.

seems...

legit?

[sulairris]

here is the code if you can read java

 

/* * To change this license header, choose License Headers in Project Properties. * To change this template file, choose Tools | Templates * and open the template in the editor. */package ccycler;import java.awt.Color;import java.awt.event.ActionEvent;import java.awt.event.ActionListener;import java.util.logging.Level;import java.util.logging.Logger;import javax.swing.*;/** * * @author nathan */public class CWin implements ActionListener {    int out=2;    int pro=0;    int st;    int fin;    int score;    boolean star=false;    JFrame window = new JFrame("Ccycler");    JPanel wall = new JPanel();    //JLabel output = new JLabel("output: "+out);    //JLabel progress = new JLabel("progress: "+pro+"/1000000");    JButton start = new JButton ("Run Benchmark");    JLabel scoreD = new JLabel("Score: "+score+" Lower Is Better.");    public CWin(){    window.setDefaultCloseOperation(window.EXIT_ON_CLOSE);    window.setVisible(true);    window.add(wall);    //wall.add(output);    //wall.add(progress);    wall.add(start);    wall.add(scoreD);    wall.setBackground(Color.red);    window.pack();    start.addActionListener((ActionListener) this);   }    public void actionPerformed(ActionEvent e){    if (e.getSource()==start){    System.out.println("test");    pro=0;    score=0;    star=true;    wall.updateUI();    scoreD.setText("Score: "+score+" Lower Is Better.");    wall.setBackground(Color.red);    wall.updateUI();        try {            Thread.sleep(3000);        } catch (InterruptedException ex) {            Logger.getLogger(CWin.class.getName()).log(Level.SEVERE, null, ex);        }    st=(int) System.currentTimeMillis()/1000;    while (pro < 1000000){        out=(int) (out/.2);        //output.setText("output: "+out);        pro=pro+1;        //progress.setText("progress: "+pro+"/1000000");        System.out.println(pro);        System.out.println(out);        wall.updateUI();        }    }    if (pro==1000000){    fin=(int) System.currentTimeMillis()/1000;    wall.setBackground(Color.green);    star=false;    score=fin-st;    System.out.println("score: "+score);    scoreD.setText("Score: "+score+" Lower Is Better.");    window.pack();    }    if (star==true){ start.setText("Benchmarking"); window.pack();}    else {start.setText("Run Benchmark"); window.pack();}    }    }

Edited April 24, 2015 by colonel_mortis
Added code tags

[Tedster]

I'm only getting a third CPU utilization with this...

And because of this low utilization:

 

4670k @ 3.4GHz

[sulairris]

I'm only getting a third CPU utilization with this...

-snip-

windows tends to do that in my testing.

-edit-

my linux box will only let it have one and a half cores.

[Tedster]

windows tends to do that in my testing.

Yeah. Then how did you get 9 out of it on a Phenom II?

[TheSLSAMG]

Yeah. Then how did you get 9 out of it on a Phenom II?

This would probably be my question, considering I scored a 57 at 4.4GHz.

[Guest]

This is not a legitimate test of CPU performance, especially if it isn't achieving 100% utilization.

[Tedster]

This would probably be my question, considering I scored a 57 at 4.4GHz.

Yeah.

[sulairris]

Yeah. Then how did you get 9 out of it on a Phenom II?

its still a WIP my laptop with an amd A6 got a 202 it looks like with the testing i have done on my systems that the OS plays a part in it.

[manon_gfx]

These kinds of tests need to be multi-threaded, and be optimized to avoid any branch mispredictions, otherwise parts of the CPU will be sitting there doing nothing.
Try spawning the same number of threads, as that there are cores/hardware threads, and do calculations doing that. That should give you a lot more precise indication of the CPU speed.

 

If you share your source code, we might be able to help you out writing something better.

EDIT:
turns out I can't read and you already uploaded your source code. You wanna get rid of all the debug print statements since those have IO delays.You don't want to update your UI on the same thread as you are bench marking, since you don't know what kind of code is behind that, and it might be doing some graphics driver stuff.

Also why do you check if pro == 1000000. It it guaranteed to be that value when you end that loop. 

NOTE: I have never written a CPU benchmark, but this is using some of my knowledge that I gathered over the years.
 

[manon_gfx]

windows tends to do that in my testing.

-edit-

my linux box will only let it have one and a half cores.

 

Linux can't give it 1.5 cores, since you program is single threaded.

[madknight3]

 

here is the code if you can read java

 

This is the programming section so yeah, many people can read Java

 

Please use code tags when posting code though.

 

its still a WIP

 

Seems like a neat project. See if you can keep improving it