Part A: - make Q1: ./vecadd00 Q2: ./vecadd01 Q3: ./matmult00 < 16 or 32 or 64> (for 256,512,1024 sized matrix) Q4: ./matmult01 < 8or 16 or 32> (for 256,512,1024 sized matrix)
Part B: Q1: g++ q1.cpp -o q1 ./q1 < K > Q2: nvcc q2.cu -o q2 ./q2 <1 or 2 or 3> (second argument is to specify which kernel to run of the three scenarios) Q3: nvcc q3.cu -o q3 ./q3 <1 or 2 or 3> (second argument is to specify which kernel to run of the three scenarios) Q4: python3 q4.py
Part C: Q1: nvcc c1.cu -o cu ./c1 Q2: nvcc c2.cu -o c2 ./c2 Q3: nvcc c3.cu -o c3 -lcudnn ./c3