A simple matrix library that use expression templates and modern concepts of C++ to solve algebric expressions. This project is built for Programming Competency Test under GSoC2O19/Boost/uBLAS.

Clang -v 7.0.1 or greater

1. Clone the repository.

git clone https://github.com/NavneetSurana/uBLAS-Programming-Competency-Test.git

2. Include lazy_matrix.h in your source file.

#include"path-to-the-cloned-repo/include/lazy_matrix.h"

3. Use clang compiler for compilation.

clang++ -std=c++17 [your src file name].cpp -o build

./build

Please refer to Simple Matrix Library for documentation.

The table below shows various operations that can be performed using this library along with short description.

Inorder to know how fast lazy_matrix libraray works I have tested it against traditional way of solving Matrix algebric expressions and the same can be found in trad_matrix.h. Using the test_case_generator.cpp file I have generated some random expression of length 300 involving operators like +,-,/,* and +=. The benchmark.h file has been used for testing and extracting the results of the test. After executing the test using main.cpp file, the results have been conveyed in the plot below. For proof one can see proof.png and for test logs one can see test_logs.txt. From the graph below one can see that Lazy Evaluation is nearly 50% more efficient than the Traditional way of Evaluation.

Note: The test above involved only element-wise operations. The operator % for Standard Matrix-Matrix Multiplication is coded without using temporaries, which unnecessarily increased the complexity due to redundant evaluation of the same elements in . One way to reduce complexity is to evaluate operator % as and when encountered and store the result in a temporary or do something related to optimization of the expression templates tree.

• Navneet Surana - Some of my works can be viewed at - GitHub
• Along with this I have also worked on solving 9X9 Sudoku Puzzle using genetic algorithm.

• Thanks to GitHub for providing such an amazing open-source platform.
• I hope this project brings in new opportunities for me.