Part of specialization from Rice University on Parallel, Concurrent and Distributed Programming
Week 1 --> Distributed PageRank in Spark
-
Ran Test for 20000 and 50000 nodes (ie Websites)
-
SpeedUp achieved -> 1.35 to 1.5
Week 2 --> A Basic File Server that responds to GET request from HTTP clients
Week 3 --> A parallel implementation of Matrix Multiply using MPI to express SPMD parallelism
-
Expected Speedup -> 0.75 * Number of Cores
-
for 8 core Laptop speedup was 6 for matrices of various sizes ranging from 800 -> 1800