Master Level Problem 3 - P1:M3 Parallelized Parser and Formula InterpreterKey Scoring Principles
Basic scoring principles used for the contest entries judging are described at the official rules page. Here is a short summary: each contest entry was scored according to the following criteria: 1) up to 100 points for solutions performance (speed); 2) up to 50 points for the solutions write-up using the 10-30-10 breakdown of points for a serial algorithm description, parallel algorithm description and performance analysis respectively; 3) a maximum of 25 bonus points for a contestants activity in the forum, calculated as 5 bonus points for each valid forum post/reply.
Input Data Sets Used for Performance Scoring
Seven different input data sets were used to compute the execution score for this problem. These data sets contain multiple formulas and expressions. Full archive of input data sets can be downloaded here.
Points in Performance Scoring
Each input data set was weighted equally and was judged individually based on a ranking scheme. The overall performance score was calculated as a sum of all seven input data sets individual points.
We allowed a total of 120 seconds execution (2 minutes) maximum for each input set; for those runs that took longer than 120 seconds or had runtime errors during execution, zero performance points were awarded. Some entries that could not be built on the MTL or were not able to complete the task correctly got zero points as well.
Successful contest entries that solved input data sets in less than 120 seconds were ranked based on their execution time and got performance points according to a reciprocal rank-weight scale. For example, the fastest solution[s] of a data set got ~14.3 points, next solution got ~7.1 points, then ~4.8 points and so on and so forth.
Execution Results and Point Spread
We received 8 contest entries in the Master Level Parallelized Parser and Formula Interpreter problem set.
Three entries successfully solved all seven input data sets. Four entries solved at least one data set. Unfortunately, one entry was incomplete and therefore unable to solve any data set.
All the timings and performance points are available in the final scoring table.
The write-up portion of each entry was read and scored by several 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.
Forum Activity and Bonus Points
Bonus points were given for contestants forum posts made before the problem entries were closed. Five points per post (maximum 25 points possible) were awarded.
Entry points and penalties.
Each contest entry got 100 entry points.
The assigned points and final score are available in the final Parallelized Parser and Formula Interpreter scoring table.
The Parallelized Parser and Formula Interpreter winners based on highest point total are the following:
These three contestants provided the solutions which correctly resolved all the input data sets and had the fastest overall code execution. They also supplied their entries with well-written, comprehensive write-ups.