Skip to content

Latest commit

 

History

History
14 lines (7 loc) · 640 Bytes

README.md

File metadata and controls

14 lines (7 loc) · 640 Bytes

Exploring Parallel T-distributed Stochastic Neighbor Embedding

Project proposal

Midterm report

Final report PDF

Final report Overleaf

Video presentation

In this project, we have successfully parallelized the Barnes-Hut t-SNE algorithm using 3 different parallelization approaches: multi-threading with OpenMP, Single Program Multiple Data with ISPC and massive parallelism with CUDA. We achieved significant speedups as compared to the baseline and analyzed the results thoroughly.