gram_schmidt

Gram Schmidt Orthogonalization on CPU/GPU

compile

nvcc -O3 -std=c++11 -D_GPU_ main.cu

performance on V100

Enter the vector length 10

Enter the # of vectors 1000

==========CPU start==========

11676 usec

==========GPU start==========

8691 usec

Enter the vector length 1000

Enter the # of vectors 100

==========CPU start==========

18185 usec

==========GPU start==========

823 usec