So I came across a few articles claiming that string concatenation using + or (concat function is slower then using a buffered approach (StringBuffer or StringBuilder). I came across this post on the topic & ran the tests on my machine.
Results I got were something like this:
On avg concat was taking 4 sec & append was taking 1 ms. ( Number of iterations- 20000).
So I kept wondering why is it so slow really ? Is it because there is excessive copying or mutliple times space allocation ! So I modified Paul's code to add some extra testing function which you can find here at pastebin. Running the results on my & my friends machine gives very very different results:
Concat Append Alloc Set Value
My machine 4000 2ms 800 80
Friends machine 8000 2ms 400 1200
Large differnce in concat times could possibly be explained due to speed of machines. But how one explains a fuzzy pattern ? We found later that he was on Java 1.6 & I was on java 1.5 . But then why would my setval function be so fast & his alloc be so slow ? What are the changes done behind the curtain ?
If anyone of you does the same test,(which is easy- copy the file & run it!) do log ur results & reply back to me...
Update: Also we thought about it later- I was on Linux & he was on windows. Is memory allocation dependent on OS too ?
Friday, August 27, 2010
Monday, August 9, 2010
P != NP !
Vinay Deolalikar, an extremely common & dull Indian name as such.
But that is what I would have said a few hours back on this name. What do I say now ? Well - this might well be the name of the most famous computer scientist of all times. Yes you read it right - of all times.
Why ? A guy with this name who works as a researcher at HP labs, Palo Alto, has claimed that he has a proof to establish that P!=NP. What it means is that if this is true, not only he takes home 1 million $, THE most difficult & important problem of computer science would be solved .
His paper is awful 103 pages long & a preliminary version can be found here.This only brings one name to my mind- Andrew Wiles ! If you don't know why- well you don't deserve to be reading this blog !
You know what is startling ? This guy was not even a computer scientist. He was an electric engineering major from IITB only. Even his PhD was in electrical engineering. Though on his homepage, he describes himself as a mathematician.
It would be a long stick in the ass of all computer scientists of this world if an inter-disciplinary researcher solves the hardest nut.
Plus its cool that he was born & brought up in Delhi & he passed out of a university which is the nearest kin to mine.(IIT Delhi).
I am so excited.
Y0 baby. C0de monkey is jumping restlessly. :)
But that is what I would have said a few hours back on this name. What do I say now ? Well - this might well be the name of the most famous computer scientist of all times. Yes you read it right - of all times.
Why ? A guy with this name who works as a researcher at HP labs, Palo Alto, has claimed that he has a proof to establish that P!=NP. What it means is that if this is true, not only he takes home 1 million $, THE most difficult & important problem of computer science would be solved .
His paper is awful 103 pages long & a preliminary version can be found here.This only brings one name to my mind- Andrew Wiles ! If you don't know why- well you don't deserve to be reading this blog !
You know what is startling ? This guy was not even a computer scientist. He was an electric engineering major from IITB only. Even his PhD was in electrical engineering. Though on his homepage, he describes himself as a mathematician.
It would be a long stick in the ass of all computer scientists of this world if an inter-disciplinary researcher solves the hardest nut.
Plus its cool that he was born & brought up in Delhi & he passed out of a university which is the nearest kin to mine.(IIT Delhi).
I am so excited.
Y0 baby. C0de monkey is jumping restlessly. :)
Subscribe to:
Posts (Atom)