Lab on "Introduction to Parallel and Distributed Computing", Spring 2018, Peking University.
-
HW 1: Data Partition and Reduction in OpenMP: Use data partition in OpenMP, parallel and for pragma, to parallelize common algorithm problems including finding the single number in an array that only appears once, computing the square of hadamard product.
-
HW 2: Task Partition and Parallelization in OpenMP: Use task partition in OpenMP, parallel, for, task and single pragma, to parallelize common algorithm problems including finding k-th smallest number in an array, finding number of inversion pairs in an array, reconstructing binary tree.
-
HW 3: Message Passing Interface for Parallel Programming: Use message passing model MPI to parallelize the process of prime number filtering.
-
- Implemented Transposed Convolution algorithm using OpenCL, measured performance with different work-item configuration.
- Utilized MPI to implement K-Nearest-Neighbor algorithm, improved node distance computation by GPU acceleration using OpenCL.