Intel® Software Network

One small test set was used to check how the applications handled a variety of intersections: co-linear segments, shared endpoints, endpoint of one segment contained in the other, and simple intersections. Three other test cases of 100000, 150000, and 200000 line segments were timed. 25 points were awarded for correct execution of the first data file, which was added to the range of 0-75 points from the last three data sets based on a linear function of where the entry's total execution time fell between the minimum time and 6 minutes (max 2 minutes per data set).

Points were deducted for identifying segment pairs that did not intersect or for missing segments that did in the output file. One point was deducted for each segment from the small test case and half a point from the number of misidentified segments in the last three data sets. This is the reason that there was not a 100 point execution score awarded.

There were 8 submissions of code solutions, 3 on Linux and 5 on Windows. There was one entry threaded with Cilk++, one with Parallel C#, and the rest in C/C++.

Point spread:

97 91
75
63 60
29
10
0

The write-up portion of each entry was read and scored by two judges. Each judge used the 10-30-10 breakdown of points for serial algorithm description, parallel algorithm description, and performance, respectively. One important component to the judging was to determine how close the submission was for publication on ISN. The assigned score was the average of the two judges scores.

Point spread:

50 50 50
33 32 31
10
25
0

Bonus points were given for contestant’s forum posts made before the problem entries were closed. Five points per post (maximum 25 points possible) were awarded.

The overall winner was BradleyKuszmaul. The fastest code execution was submitted by akki. The judges were most impressed with the write-up submissions that accompanied the entries, especially those from Dmitry Vyukov, akki, BradleyKuszmaul who all had excellent write-up submissions.