Bioinformatics Benchmarks Project

GSoC
1

Hello mentors!
I am a 4th year Electrical and Computer Engineering student at the University of Thessaly, Volos, Greece and I am interested in Bioinformatics Benchmarks project. I have started learning Chapel recently and implemented the suggested "word count in strings" program.

I would like to hear your opinion on my approach.
In addition, I have already submitted a draft proposal through the GSoC website and I am waiting for possible feedback.

Thanks, I hope to meet some of you soon.
Antonis

2

Hi Antonis,

Quick thoughts:

  • Why not wordMap[word] += 1? See as an example:
use Map;

var m = new map(string,int);

m["foo"] += 1;
m["foo"] += 1;
m["foo"] += 1;

writeln(m);
  • There might be some performance implications of the idiom you use. And the best way to test different versions is to have a much larger string that has reoccurring words and run both idioms on the same string. Make sure you compile your applications with the --fast flag. You can use the timers in the Time module to measure time: Time — Chapel Documentation 1.25

Engin

3

Hi Engin,
I didn't know that you can access an element of a map like that. I ran some quick performance tests for large strings with repetitive words and indeed your suggestion performs two times faster.

Thank you for your interest,
Antonis