• Next Thread Prev Thread

• Posts   Search Threads

• Clay Breshears (Intel) November 12, 2009 8:57 AM PST    Scoring Criteria Five different sets of data files were used to compute the execution score for this problem: tenCube.txt was all points where the absolute value of each coordinate is in [1,10]. The solution is all points with at least one "+/-10" coordinate (labeled with "M" as first character). tenCorner.txt was the tenCube.txt data set with the eight points (+/-11, +/-11, +/-11) added. The solution is the eight new points (labeled as "PPPx" with x in [1,8]). 85Shell.txt was a set of points that were all located 85 units from the origin. The solution is all points. Two files of random points within the space. The count of points was 10,000 and 15,000. A proportional amount of penalty points were deducted for each data set that was not completely correct. 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. 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.  (Looks like he got enough of the corner cases covered.)  Attachments  3DConvexHullData.zip Clay Breshears (Intel) November 12, 2009 11:30 AM PST Rate   #1 Re: Scoring Criteria One other thing, during the 20 OCT episode of Parallel Programming Talk (around 2:10 - 3:07), I noted that there is a journal article published in the Communications of the ACM (volume 20, 1977, pp.87-93) that gave details about 3-D Convex Hull algorithm.  This article is available online at the CACM site.  (I'm a member of ACM, so I'm not sure if someone that isn't a member can get the same access).  If you're interested try this link:  http://portal.acm.org/citation.cfm?id=359430&dl=GUIDE&coll=GUIDE&CFID=62641943&CFTOKEN=32038911. In the PPT show, I also noted that there is a TBB parallel version of (2-D) Convex Hull that is in the examples directory. --clay avparate November 12, 2009 9:02 PM PST Rate   #2 Reply To #1 Re: Scoring Criteria Quoting - Clay Breshears (Intel) One other thing, during the 20 OCT episode of Parallel Programming Talk (around 2:10 - 3:07), I noted that there is a journal article published in the Communications of the ACM (volume 20, 1977, pp.87-93) that gave details about 3-D Convex Hull algorithm.  This article is available online at the CACM site.  (I'm a member of ACM, so I'm not sure if someone that isn't a member can get the same access).  If you're interested try this link:  http://portal.acm.org/citation.cfm?id=359430&dl=GUIDE&coll=GUIDE&CFID=62641943&CFTOKEN=32038911. In the PPT show, I also noted that there is a TBB parallel version of (2-D) Convex Hull that is in the examples directory. --clay Congrats BradleyKuszmaul... akki November 13, 2009 10:48 AM PST Rate   #3 Re: Scoring Criteria Quoting - Clay Breshears (Intel) Five different sets of data files were used to compute the execution score for this problem: tenCube.txt was all points where the absolute value of each coordinate is in [1,10]. The solution is all points with at least one "+/-10" coordinate (labeled with "M" as first character). tenCorner.txt was the tenCube.txt data set with the eight points (+/-11, +/-11, +/-11) added. The solution is the eight new points (labeled as "PPPx" with x in [1,8]). 85Shell.txt was a set of points that were all located 85 units from the origin. The solution is all points. Two files of random points within the space. The count of points was 10,000 and 15,000. A proportional amount of penalty points were deducted for each data set that was not completely correct. 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. 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.  (Looks like he got enough of the corner cases covered.) Congratulations Bradley! 邓辉 November 16, 2009 6:48 AM PST Rate   #4 Reply To #3 Re: Scoring Criteria Congratulations Bradley! Too busy working. Did not have time to race. Pity 写字楼里写字间,写字间里程序员 程序人员写程序,又拿程序换酒钱 酒醒只在网上坐,酒醉还来网下眠 酒醉酒醒日复日,网上网下年复年

• = Best Answer

  Next Thread Prev Thread

Intel Software Network Forums Statistics

16,371 users have contributed to 46,346 threads and 163,981 posts to date.

For more complete information about compiler optimizations, see our Optimization Notice.