/All-Pair-Shortest-Path-CUDA

Accelerate algorithm block all-pair shortest path through CUDA

Primary LanguageCudaMIT LicenseMIT

All Pair Shortest Path parallel in CUDA

Implement parallel OAll Pair Shortest Path with MPI and OpenMP.

Experiment System Spec

Clusters:

  • node 1:

    • Intel Xeon CPU L5640 @2.27GHz (2x6-cores)
    • NVIDIA Tesla K20m (5G GDDR5) x2

  • node2 and node3:

    • Intel Xeon CPU X5670 @2.93GHz (2x6-cores)
    • NVIDIA Tesla M2090 (6G GDDR5) x2

  • Network: connectd by 1000M Gigabit Ethernet

Report