Assignment Tutorial
MSBD5009/COMP5112 Parallel Programming Assignment 1: Super-mer Generation with MPI
Tutorial Overview
Copyright By PowCoder代写 加微信 powcoder
• Problem Description
• Implementation Instruction • Environment Setup
Problem Description
• Basic Concepts 1. Read
• A DNA fragment with base A, C, T, G (i.e., a string contains ‘A’, ‘C’, ‘T’, ‘G’ only).
• CAAATTACTGCATA 2. K-mer
• A length-𝑘 substring on a read. A read of length 𝑛 contains 𝑛 − 𝑘 + 1 k-mers.
• (k=9)CAAATTACT,AAATTACTG,…, TACTGCATA are the k-mers of the above read
3. Minimizer 4. Super-mer
Problem Description
• Basic Concepts 1. Read
3. Minimizer
• The lexicographically smallest length-𝑝 substring of a k-mer.
• (p=5)TheminimizerofCAAATTACTisAAATT. 4. Super-mer
• A substring of a read generated by merging multiple consecutive k-mers which have the same minimizer value.
• (k=9, p=5) The first super-mer in the read CAAATTACTGCATA will be CAAATTACTG because the first two k-mers have the same minimizer AAATT.
Problem Description
• Your Task
• Many reads
• Given in CSR format in the program
• All the super-mers generated from these reads
• You need to save all the super-mers to a vector of strings “all_supermers” in Process 0
Implementations
• The code skeleton gensuper-mer_mpi.cpp
• Already implemented:
• MPI initialization and finalization
• Loading reads from the dataset file and converting to CSR format
• Result correctness checking
• Outputting super-mers to text file
• * Function read2supermers(…) which can convert a read to its corresponding super- mers
• You need to:
• Scatter the read data to each MPI process
• Perform the super-mer generation in each process
• You can refer to the sequential version to know the usage of the function read2supermers(…) • Gather all the super-mers to Process 0 and store in the vector “all_supermers“
• Each string represents a super-mer
• The order in the vector doesn’t matter
Implementations
• Only write your code in the specified area of gensuper- mer_mpi.cpp and only submit this file to Canvas.
程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com