Workloads
Workloads Overview
During the competition, each team must run a variety of workloads.
Each team will be judged on their ability to run the workloads efficiently and accurately.
HPCG Benchmark
High Performance Conjugate Gradient (HPCG) is an industry standard benchmark designed to profile HPC systems.
HPCG Benchmark will be used to measure the raw compute performance of your system. This benchmark is used as a metric to rank the TOP500 computer systems in the world.
Median of Merged Array
Problem Description
Given two unsorted arrays A and B of size m and n (respectively), your task to find the median value of the sorted array, obtained by the merge of input arrays A and B.
Note:
- If the merged array has an even number of elements (i.e., is_even(m+n) == True), the median is the average between the two center elements.
- Two arrays will be provided in separate text files. Each element of the array is separated by a new line.
Test Datasets
Download Median DatasetAdditional Resources
Standardizing Recipe Ingredients
Background
The simplified-recipes-1M Dataset by Dominik Schmidt contains approximately 1,000,000 carefully cleaned and preprocessed recipes. The data comes from five different base datasets which were merged in order to create a more complete recipe collection. All recipes have been meticulously cleaned and standardized.
During the preprocessing step, Dominik faced a challenge: he needed to standardize the receipe ingredients.
Problem Description
Your task is to count the top 20 most used ingredients in an example dataset. You will be provided with a text file that lists recipe IDs and a string of ingredients. You must then match the substrings against the dictionary of known ingredients and present the top 20 most used ingredients in our dataset.
Test Datasets
Download Recipe DatasetDrilling Circuit Boards
Background
Printed circuit boards (or PCB) are used by current generation electronic devices. Each PCB can have hundreds to thousands of metalized holes called vias. During the manufacturing process, each hole must be drilled once. Finding the most optimal way for an automatic drilling machine to visit each via can significantly speed-up PCB production time.
Problem Description
Given the positions (x,y) of all of these vias, your task is to find the shortest tour for our drilling machine to visit all the vias once.
Note:
- Our drilling machine can move diagonally. Thus, the euclidean distance metric should be used.
- The final result should be within 1% of the optimal solution
Test Datasets
Download Circuit Board DatasetAdditional Resources
Mystery Application
At the start of the competition, a new workload will be released. Competitiors will be required to code, compile, and run this workload during the competition along with the above workloads.
Contact
Questions? Email neu.hpcboston@gmail.com
About NUCAR
HPC Boston is being run by students from Northeastern University’s Computer Architecture Laboratory, also known as NUCAR. NUCAR conducts research in areas such as computer security, GPU computing, and embedded systems. The lab is overseen by Professor David Kaeli.