Differences between revisions 9 and 10
 ⇤ ← Revision 9 as of 2011-02-15 05:56:05 → Size: 2182 Editor: edwardyoon Comment: ← Revision 10 as of 2011-02-15 23:56:08 → ⇥ Size: 1967 Editor: edwardyoon Comment: Deletions are marked like this. Additions are marked like this. Line 3: Line 3: The value of PI can be calculated in a number of ways. Consider the following method of estimating PI The value of PI can be calculated in a number of ways. In this example, we are estimating PI using following way: Line 5: Line 5: * Inscribe a circle in a square * Randomly generate points in the square * Determine the number of points in the square that are also in the circle * Let r be the number of points in the circle divided by the number of points in the square * PI ~ 4 r * Each task executes locally its portion of the loop a number of times. Line 11: Line 7: Serial pseudo code for this procedure as below: {{{   iterations = 10000  circle_count = 0 Line 13: Line 11: {{{iterations = 10000circle_count = 0do j = 1,iterations do j = 1,iterations Line 23: Line 17: end do end do Line 25: Line 19: PI = 4.0*circle_count/iterations PI = 4.0*circle_count/iterations Line 28: Line 22: == The BSP implementation for Pi == * One task acts as master and collects the results through the BSP communication interface. Line 30: Line 24: A distributed strategy in HAMA with [[http://wiki.apache.org/hama/Architecture#BSP_Programming_Model|BSP programming model]], is break the loop into portions that can be executed by the tasks. {{{  PI = pi_sum / n_processes}}} Line 32: Line 28: * Each task executes locally its portion of the loop a number of times.  * One task acts as master and collects the results through the BSP communication interface. 1) Each process computes the value of Pi locally, and 2) sends it to master task using send() function. Then, 3) the master task can recieve the messages using sync() function. Finally, we can calculate the average value of sum of PI values from each peers as below:

## Pi Estimator

The value of PI can be calculated in a number of ways. In this example, we are estimating PI using following way:

• Each task executes locally its portion of the loop a number of times.

```  iterations = 10000
circle_count = 0

do j = 1,iterations
generate 2 random numbers between 0 and 1
xcoordinate = random1
ycoordinate = random2
if (xcoordinate, ycoordinate) inside circle
then circle_count = circle_count + 1
end do

PI = 4.0*circle_count/iterations```
• One task acts as master and collects the results through the BSP communication interface.

`  PI = pi_sum / n_processes`

1) Each process computes the value of Pi locally, and 2) sends it to master task using send() function. Then, 3) the master task can recieve the messages using sync() function. Finally, we can calculate the average value of sum of PI values from each peers as below:

```    public void bsp(BSPPeerProtocol bspPeer) throws IOException,
KeeperException, InterruptedException {
int in = 0, out = 0;
for (int i = 0; i < iterations; i++) {
double x = 2.0 * Math.random() - 1.0, y = 2.0 * Math.random() - 1.0;
if ((Math.sqrt(x * x + y * y) < 1.0)) {
in++;
} else {
out++;
}
}

byte[] tagName = Bytes.toBytes(bspPeer.getPeerName());
byte[] myData = Bytes.toBytes(4.0 * (double) in / (double) iterations);
BSPMessage estimate = new BSPMessage(tagName, myData);

bspPeer.sync();

double pi = 0.0;
int numPeers = bspPeer.getNumCurrentMessages();