How I kicked off GSoC

Zero to hero

What Prompted me??

I started my third year thinking I should do something that would put me different from the rest and one of my professors suggested me as to why don’t I apply for GSoC. I don’t know why but I took the suggestion rather seriously, thanks to the bet I had with one of my friend(who is about to complete his MBBS) that whoever earns first will buy the other a “RayBan shades”. Well, that’s it. I was determined. I started my research early, probably during the start of February(I knew I want to buy my friend, his shades and also buy mine too, in the process).

What experiences I had before??

I started looking at previous years’ GSoC projects(having had little experience with Open Source)  and started learning how to contribute. I was also very fascinated to the amount of knowledge one could gain just by googling and browsing web pages . I discovered very soon, as to what an immensely great tool , email, through which I could chat with anyone in the open source world and ask seemingly stupid questions and always expect to get a gentle reply back with an answer. Well, that held me spell bound and I knew I want to contribute to Open Source.

How did I begin??

About the middle of March, I discovered that my passion for Python as a programming language increased , after understanding how easy it is as a language. Added to that, my popularity among my fellow classmates increased when I started evangelizing Python(thanks to my seniors for introducing it, I guess I did a decent job popularizing the language). And I started contributing to PSF(Python Software Foundation) , started with a simple bug to fix documentation and slowly my interactivity in IRC increased and I started liking one of the project one of the community member proposed.

A twist in the story??

There I was, still a noob and not knowing how to convince my probable mentor that I could complete the project, given direction. About this juncture, a fellow student(from some university in France) mailed this particular mentor that he was interested in the project . Do, remember, I was part of the mailing list and follow the happenings of it. So, I was furious knowing that I had a competition(having put so much effort) and I was not willing to compromise my project (knowing that this is the one project I actually understood and started researching a little bit too). The other projects require me to have some domain knowledge. I went back to my teachers, seniors, friends and Google and started asking the question , “how would i solve the problem the mentor posted?” . I framed a couple of answers, though very noobish , but at least I could reply the email thread posting my understanding of the problem and how I would solve it and also ask various questions I had in my mind. Well, the mentor replied, immediately to my surprise, and responded back with comments as well as answers to the questions I posed. Again, my nemesis/competitor replied back(he having good knowledge about the problem domain). I knew it was not going to be easy. Hence, I went back again, through all my sources, made further understanding of the problem and posted back again. I guess, about 20 mails in the thread , till we(all three of us) decided we should catch up in IRC and discuss more.

The conclusion:

Well, at IRC , most of senior members from the community were present, and they suggested that they should probably extend the scope of the project(since two students were interested in one project and showed immense passion). Unsurprisingly, over multiple meetings, the project scope was expanded, both the students were given equal important but independent tasks and both the students got opportunity to say they are Google Summer of Code students. Thank goodness, we decided to built the project from scratch giving us more than enough work on our plate.

Movie titles:

1) In the open source world, there is no competition , it is only “COLLABORATION“.

2) Why give up, when you can win??

3) From Zero to Hero!!

4) A prodigy in making

p.s. I still owe my friend his shades . *sshole, I am still waiting for him to visit me so that I can buy him his  shades and buy mine too.  Also, I know its been two years since the story happened, but it is never too late to share, don’t you agree??

Advertisements

Calculation of Pi

After my initial encountering with OpenCL, I wanted to explore more. So I wrote a small program to calculate Pi .
This uses a very simple algorithm to calculate Pi

  1. Inscribe a circle in a square
  2. Randomly generate points in the square
  3. Determine the number of points in the square that are also in the circle
  4. Let r be the number of points in the circle divided by the number of points in the square
  5. PI ~ 4 r
  6. Note that the more points generated, the better the approximation

This gave the value of Pi as 3.141172 . Pretty decent, I suppose 🙂
 

The code

 

import pyopencl as cl 
import numpy
import random

class circle(object):
    def __init__(self, r=100):
        self.r = r
    def inside_the_circle(self, x, y):
        if x**2+y**2 <= r**2: 
           return 1
        else:
           return 0

class CL:
    def __init__(self, size=1000):
        self.size = size
        self.ctx = cl.create_some_context()
        self.queue = cl.CommandQueue(self.ctx)

    def load_program(self):
        f = """
        __kernel void picalc(__global float* a, __global float* b, __global float* c)
        {
         unsigned int i = get_global_id(0);

        if (a[i]*a[i]+ b[i]*b[i] < 100*100)
           {
            c[i] = 1;
           }
        else
           {
            c[i]=0;
           }
        }
        """
        self.program = cl.Program(self.ctx, f).build()

    def popcorn(self):
        mf = cl.mem_flags
        x = [random.uniform(-100,100) for i in range(self.size)]

        y = [random.uniform(-100,100) for i in range(self.size)]

        self.a = numpy.array(x, dtype=numpy.float32)
        self.b = numpy.array(y, dtype=numpy.float32)

        self.a_buf = cl.Buffer(self.ctx, mf.READ_ONLY | mf.COPY_HOST_PTR, hostbuf=self.a)
        self.b_buf = cl.Buffer(self.ctx, mf.READ_ONLY | mf.COPY_HOST_PTR, hostbuf=self.b)
        self.dest_buf = cl.Buffer(self.ctx, mf.WRITE_ONLY, self.b.nbytes)

    def execute(self):
        self.program.picalc(self.queue, self.a.shape, None, self.a_buf, 
                            self.b_buf, self.dest_buf)
        self.c = numpy.empty_like(self.a)
        cl.enqueue_read_buffer(self.queue, self.dest_buf, self.c).wait()
        
    def calculate_pi(self):
        number_in_circle = 0
        for i in self.c:
            number_in_circle = number_in_circle + i
        pi = number_in_circle*4 / self.size
        print 'pi = ', pi



if __name__ == '__main__':
    ex = CL(1000000)
    ex.load_program()
    ex.popcorn()
    ex.execute()
    ex.calculate_pi()

 

CPU vs GPU performance comparision with OpenCL

I recently had opportunity to explore an awesome library called OpenCL (Open Computing Language) which enables me to create programs which helps me utilize the computation power of my Graphic Card. I wanted to try out how much faster a normal program (addition of elements to two arrays) would work if I parallize the program using OpenCL.

Source
Using OpenCL

import pyopencl as cl
import numpy
import sys

class CL(object):
    def __init__(self, size=10):
        self.size = size
        self.ctx = cl.create_some_context()
        self.queue = cl.CommandQueue(self.ctx)

    def load_program(self):
        fstr="""
		__kernel void part1(__global float* a, __global float* b, __global float* c)
		{
       		unsigned int i = get_global_id(0);

	       c[i] = a[i] + b[i];
		}
	     """
        self.program = cl.Program(self.ctx, fstr).build()

    def popCorn(self):
        mf = cl.mem_flags

        self.a = numpy.array(range(self.size), dtype=numpy.float128)
        self.b = numpy.array(range(self.size), dtype=numpy.float128)

        self.a_buf = cl.Buffer(self.ctx, mf.READ_ONLY | mf.COPY_HOST_PTR,
                               hostbuf=self.a)
        self.b_buf = cl.Buffer(self.ctx, mf.READ_ONLY | mf.COPY_HOST_PTR,
                               hostbuf=self.b)
        self.dest_buf = cl.Buffer(self.ctx, mf.WRITE_ONLY, self.b.nbytes)

    def execute(self):
        self.program.part1(self.queue, self.a.shape, None, self.a_buf, self.b_buf, self.dest_buf)
        c = numpy.empty_like(self.a)
        cl.enqueue_read_buffer(self.queue, self.dest_buf, c).wait()
        print "a", self.a
        print "b", self.b
        print "c", c

if __name__ == '__main__':
    matrixmul = CL(10000000)
    matrixmul.load_program()
    matrixmul.popCorn()
    matrixmul.execute()

Normal program without Optimization

def add(size=10):
    a = tuple([float(i) for i in range(size)])
    b = tuple([float(j) for j in range(size)])
    c = [None for i in range(size)]
    for i in range(size):
        c[i] = a[i]+b[i]

    #print "a", a
    #print "b", b
    print "c", c[:1000]

add(1000000)

I compared the performance of both the programs using the tool “time” available in Linux and I noted down the “sys time”
Heres the comparision:

 

Size Using GPU Without GPU ( i.e. CPU)
100 0.130s 0.030s
1000 0.100s 0.010s
10000 0.130s 0.010s
100000 0.150s 0.050s
1000000 0.170s 0.150s
10000000 0.600s 1.150s

 
Cleary you see that the GPU outperforms CPU at higher values of size as the program is able to use multiple threads provided by the GPU. At lower values of size there is an appreciable access time associated with GPU, so CPU performs faster.

Pycon India

My first visit to any Pycon Conference(this time at Pune, India) happened last week. It was really great( as expected !!! ) . Incidentally I gave a talk there too, on PyTI (Python Testing Infrastructure) . I got to make few more friends and learnt some interesting stuff out there. One of the notable talk to be watched is definitely the keynote by Raymond Hettinger (again no surprises here as keynote is the soul of any conference ) . I had a great time in Pune and also Mumbai which I visited on sunday 🙂