Algorithms Complexity - A brief analisys on computational cost

Once upon a time where computer scientists worries about computational resources limitation. They needed to work on the available hardware that wasn't that fast compared to nowadays. This presentation is about how we can back to the golden age of developers where performance wasn't optimized by buying expensive hardwares.

Transcript

1. Algorithms Complexity A brief analisys to computational cost
 
 Emanuel de Souza
2. Numbers and Facts 31 years old++ (and counting)
 22 years programming++ (first code at 9 years old)
 13 years professional exp++ (nonstop coder)
 2 daughters (not planning incrementing this counter)
 1 wife (neither this one)
 
 CTO at Contentools Inc
 Python evangelist
 PyBr Member
 Simplicity lover
 Open source contributor
3. Computational Cost
 What involves it?
4. Memory CPU Bandwidth GPU Storage
5. Memory CPU Bandwidth GPU Storage
6. How can we spend less computational resources?
7. Optimization!
8. What is complexity?
10. Algorithm Complexity • Two dimensions (time and space) • Time = CPU • Space = Memory • Classified using Big-O Notation
11. Big-O Complexity
12. Given a number “x" calculate the result of sum of terms from “1" given 1 < x.
 e.g: 1 + 2 + 3 + … + 10 =
13. The naive O(nˆ2) solution For each value from 1 to x, implements a counter to be increased CHECK THE CODE AT
 http://github.com/emanuelcds/complexity
14. The fair O(n) solution For each value from 1 to x, increase the result by the current iterator value CHECK THE CODE AT
 http://github.com/emanuelcds/complexity
15. The optimized O(1) solution Writing the sequence 1, 2, …, n and repeat the same sequence on inverted order underneath each other CHECK THE CODE AT
 http://github.com/emanuelcds/complexity If you just sum each column, you will have the same result So, we can easily deduce a single equation for the result, obtaining a O(1) observation
16. "We are paid to think before coding any crap!”
 - Machado de Assis
17. “Don't follow-me, I don’t have a clue!”
 Thank you! github.com/emanuelcds emanuelcds