Open MPI User's Mailing List Archives

From: Götz Waschk (goetz.waschk_at_[hidden])
Date: 2007-03-28 10:25:01

• Next message: George Bosilca: "Re: [OMPI users] Measuring MPI message size used by application"
• Previous message: stephen mulcahy: "[OMPI users] Measuring MPI message size used by application"
• In reply to: Warner Yuen: "[OMPI users] Odd behavior with slots=4"
• Next in thread: George Bosilca: "Re: [OMPI users] Odd behavior with slots=4"

On 3/28/07, Warner Yuen <wyuen_at_[hidden]> wrote:
> variation 1: hostname
> real 0m35.391s
>
> variation 2: hostname slots=4
> real 0m45.698s
>
> variation 3: hostname slots=2
> real 0m38.761s

Hi Warner,

how many nodes does your cluster have? I assume it is using only one
process per node by default. With slots=4 you might hit the Xeon
bottle neck.

Regards, Götz Waschk

-- 
AL I:40: Do what thou wilt shall be the whole of the Law.

• Next message: George Bosilca: "Re: [OMPI users] Measuring MPI message size used by application"
• Previous message: stephen mulcahy: "[OMPI users] Measuring MPI message size used by application"
• In reply to: Warner Yuen: "[OMPI users] Odd behavior with slots=4"
• Next in thread: George Bosilca: "Re: [OMPI users] Odd behavior with slots=4"