From: Jean-Paul Dzisiak Newsgroups: comp.parallel.mpi Subject: Re: performance of MPICH 1.1 under AIX 4.3 Date: Mon, 01 Mar 1999 09:44:44 +0100 Organization: Environment Center - ULg Message-Id: <36DA537B.E6CB713C@ulg.ac.be> References: <36D6C4F7.BE30B4B4@ulg.ac.be> Mime-Version: 1.0 Content-Type: multipart/alternative; boundary="------------5D2C28FC9CB07F25A4CDAFDB" Xref: ukc comp.parallel.mpi:4682 --------------5D2C28FC9CB07F25A4CDAFDB Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable Jean-Paul Dzisiak wrote: > Hello ! > > I've 3 processes running on an AIX 4.3 IBM workstation. I'm using the > "p4" device: > > [process 1 (P1)] > || > [process 2 (P2)] <=3D=3D=3D> [process 3 (P3)] > > All the messages use MPI_SEND / MPI_RECV > > Per iteration, I have: > i) P1 sends P2 600 blocks of 30 reals > ii) P2 sends P3 1 block of 300 reals > iii) P3 sends P1 1 block of 130 reals > > When I don't activate messages P3 to P2 [ P2 =3D> P3], my whole > enviroment run very fast. OK. > but when I activate messages P3 to P2 (even a single REAL !) [P2 <=3D> > P3], performances collapse > > Why a single MPI_SEND/MPI_RECV message (that send only 130 real !) > significantly slows down > > the processes ? Is there a known bug ? Do I have to update MPICH to 1.2= ? > After 217 iterations, I have the following error message: =3D(MOHICAN)=3D envoi des percolats... 217 =3D(MOHICAN)=3D percolats envoyes =3D(MOHICAN)=3D reception des percolats... 217 0 - Could not dynamically allocate internal handle : Internal MPI error! [0] Aborting program ! [0] Aborting program! p0_8474: p4_error: : 16 bm_list_17388: p4_error: interrupt SIGINT: 2 p3_21092: p4_error: interrupt SIGINT: 2 rm_l_3_19640: p4_error: interrupt SIGINT: 2 p2_22876: p4_error: interrupt SIGINT: 2 rm_l_1_21372: p4_error: interrupt SIGINT: 2 p1_23200: p4_error: interrupt SIGINT: 2 rm_l_2_20170: p4_error: interrupt SIGINT: 2 P4 procgroup file is /home/cenv/jep/ceme/mohican/master/config/procgroup.= Thanks for your comments ... -- Jean-Paul Dzisiak CEME - Universit=E9 de Li=E8ge Centre d'Etude et de Mod=E9lisation de l'Environnement Sart-Tilman B5 B-4000 Li=E8ge BELGIQUE T=E9l +32 (0) 4 366 23 56 Fax +32 (0) 4 366 23 55 e-mail JP.Dzisiak@ulg.ac.be --------------5D2C28FC9CB07F25A4CDAFDB Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit Jean-Paul Dzisiak wrote:
 Hello !

I've 3 processes running on an AIX 4.3 IBM workstation. I'm using the "p4" device:

[process 1 (P1)]
   ||
[process 2 (P2)]  <===> [process 3 (P3)]

All the messages use MPI_SEND / MPI_RECV

Per iteration, I have:
  i)   P1 sends P2 600 blocks of 30 reals
  ii)  P2 sends P3 1 block of 300 reals
  iii) P3 sends P1 1 block of 130 reals

When I don't activate messages P3 to P2 [ P2 => P3], my whole enviroment run very fast. OK.
but when I activate messages P3 to P2 (even a single REAL !) [P2 <=> P3], performances collapse

Why a single MPI_SEND/MPI_RECV message (that send only 130 real !) significantly slows down 

the processes ? Is there a known bug ? Do I have to update MPICH to 1.2 ?



After 217 iterations, I have the following error message:



 =(MOHICAN)= envoi des percolats... 217
 =(MOHICAN)= percolats envoyes
 =(MOHICAN)= reception des percolats... 217
0 - Could not dynamically allocate internal handle : Internal MPI error!
[0]  Aborting program !
[0] Aborting program!
p0_8474:  p4_error: : 16
bm_list_17388:  p4_error: interrupt SIGINT: 2
p3_21092:  p4_error: interrupt SIGINT: 2
rm_l_3_19640:  p4_error: interrupt SIGINT: 2
p2_22876:  p4_error: interrupt SIGINT: 2
rm_l_1_21372:  p4_error: interrupt SIGINT: 2
p1_23200:  p4_error: interrupt SIGINT: 2
rm_l_2_20170:  p4_error: interrupt SIGINT: 2
P4 procgroup file is /home/cenv/jep/ceme/mohican/master/config/procgroup.




Thanks for your comments ...



-- 
Jean-Paul Dzisiak

CEME - Université de Liège
Centre d'Etude et de Modélisation de l'Environnement
Sart-Tilman B5
B-4000 Liège
BELGIQUE

Tél +32 (0) 4 366 23 56
Fax +32 (0) 4 366 23 55
e-mail JP.Dzisiak@ulg.ac.be

 

--------------5D2C28FC9CB07F25A4CDAFDB--