Skip to content

Releases: parthsl/TSP_XeonPhi

Optimization of 2opt based Algorithm for multiple core systems

08 May 05:40
Compare
Choose a tag to compare

The Algorithm for solving TSP using 2opt heuristics.

The code optimized for IBM POWER8 and Intel Xeon Phi: Knights Landing systems


Features:

  1. Compatible with and multi-core and many-core architectures.
  2. Fully vectorized for both the systems with AVX2 and AVX512 technology.
  3. Uses OpenMP for multi-threading on a single node.
  4. Compatible with multi-nodes using MPI based implementation.
  5. Almost linear scaling across a cluster.
  6. Lower memory footprint compared to other algorithms.