LAM/MPI General User's Mailing List Archives

From: madhuri gupta (madhurisd_at_[hidden])
Date: 2009-10-29 05:01:54

• Next message: Elias SantAna: "Re: LAM: lam Digest, Vol 1286, Issue 1"
• Previous message: Natarajan CS: "LAM: Passing dynamic array as whole?"
• Next in thread: Eugene Loh: "Re: LAM: execution time of a parallel program"
• Reply: Eugene Loh: "Re: LAM: execution time of a parallel program"

I m working on a shared linux cluster with MPI. I m optimizing the execution time of my algorithm.So i want to measure the exact execution time of the program.How can i do this????????
     The other thing i want to know is that how can i use MPI_Gather to gather 1D arrays from all processes to a 2D array on the root process. Actually using send & recv will be too costly from communication point of view.

Madhuri
    Research Scholar
    Department of Mathematics
    IIT Roorkee

• text/html attachment: attachment

• Next message: Elias SantAna: "Re: LAM: lam Digest, Vol 1286, Issue 1"
• Previous message: Natarajan CS: "LAM: Passing dynamic array as whole?"
• Next in thread: Eugene Loh: "Re: LAM: execution time of a parallel program"
• Reply: Eugene Loh: "Re: LAM: execution time of a parallel program"