Tracing Data Movements within MPI Collectives | Zendy

Kevin Brown | Zendy; Jens Domke | Zendy; Satoshi Matsuoka | Zendy

AI Assistant Blog Pricing

Home ZAIA Blog

Open Access

Tracing Data Movements within MPI Collectives

Author(s) -

Kevin Brown,

Jens Domke,

Satoshi Matsuoka

Publication year - 2014

Publication title -

tokyo tech research repository (tokyo institute of technology)

Language(s) - English

Resource type - Conference proceedings

DOI - 10.1145/2642769.2642789

Subject(s) - infiniband , computer science , benchmark (surveying) , tracing , visualization , overhead (engineering) , trace (psycholinguistics) , message passing , data visualization , message passing interface , footprint , distributed computing , parallel computing , operating system , data mining , linguistics , philosophy , paleontology , geodesy , biology , geography

We propose extending common performance measurement and visualization tools to identify network bottlenecks within MPI collectives. By creating additional trace points in the Peruse utility of Open MPI, we track low-level InfiniBand communication events and then visualize the communication profile in Boxfish for a more comprehensive analysis. The proposed tool-chain is non-intrusive and incurs less than 0.1% runtime overhead with the NPB FT benchmark.

The content you want is available to Zendy users.

Already have an account? Click here to sign in.

Having issues? You can contact us here

Accelerating Research