LAM: problem with mpiexec, Fink & Leopard

Hi,

I'm new to this list and to lam-mpi, so forgive me if I'm
asking a common 
question.  I tried to find the answer on the FAQs and
archives, but either 
didn't find it, or didn't know it when I saw it.

I've installed lam-mpi on my MacBook Pro, OS 10.5, using
Fink.  I'm trying 
to use lam-mpi to run the downloadable cosmology code
Gadget2.  When I run 
the program using mpiexec, I get the following output:

05:00 PM btpro:Gadget2> lamboot -v

LAM 7.1.3/MPI 2 C++/ROMIO - Indiana University

n-1<85442> ssi:boot:base:linear: booting n0
(localhost)
n-1<85442> ssi:boot:base:linear: finished
05:01 PM btpro:Gadget2>
lamnodesn0	localhost:1:origin,this_node
05:01 PM btpro:Gadget2> lam-mpiexec -np 1 ./Gadget2 
../bobtodd_tests/thick_disk_test01/thick_disk01_paramfiles/t
hick_disk01.param
------------------------------------------------------------
--------------
Failed to find the following executable:

Host:       Macintosh.local
Executable: lam_appschema_ysCT98

Cannot continue.
------------------------------------------------------------
--------------
05:01 PM btpro:Gadget2>

I have not yet been able to figure out what this problem is.
 Any help 
would be greatly appreciated.  (I recently figured out that,
yes, I should 
be using OpenMPI, bundled with 10.5... .  I'll cross that
bridge once I 
get this ironed out. 

Thanks very much,
Todd

PS -- output of laminfo follows.

02:46 PM btpro:Gadget2> laminfo
              LAM/MPI: 7.1.3
               Prefix: /sw
         Architecture: i386-apple-darwin9.1.0
        Configured by: root
        Configured on: Thu Feb  7 14:48:44 CST 2008
       Configure host: Macintosh.local
       Memory manager: darwin7malloc
           C bindings: yes
         C++ bindings: yes
     Fortran bindings: yes
           C compiler: gcc
         C++ compiler: g++
     Fortran compiler: /sw/bin/gfortran
      Fortran symbols: underscore
          C profiling: yes
        C++ profiling: yes
    Fortran profiling: yes
       C++ exceptions: no
       Thread support: yes
        ROMIO support: yes
         IMPI support: no
        Debug support: no
         Purify clean: no
             SSI boot: globus (API v1.1, Module v0.6)
             SSI boot: rsh (API v1.1, Module v1.1)
             SSI boot: slurm (API v1.1, Module v1.0)
             SSI coll: lam_basic (API v1.1, Module v7.1)
             SSI coll: shmem (API v1.1, Module v1.0)
             SSI coll: smp (API v1.1, Module v1.2)
              SSI rpi: crtcp (API v1.1, Module v1.1)
              SSI rpi: lamd (API v1.0, Module v7.1)
              SSI rpi: sysv (API v1.0, Module v7.1)
              SSI rpi: tcp (API v1.0, Module v7.1)
              SSI rpi: usysv (API v1.0, Module v7.1)
               SSI cr: self (API v1.0, Module v1.0)

_______________________________________________
This list is archived at http://www.l
am-mpi.org/MailArchives/lam/

Hi Todd,

Better switch to OpenMPI as soon as possible.
I spent couple of days installing and then trying to run LAM
7.1.2 and  
7.1.4 on the same computer and OS you have.
I managed to install LAM, then I was able to compile my
parallel codes  
BUT there was no way to run them.
Strange errors were reported and finally I gave up.

Jeff Squyres suggested to go for OpenMPI. I did it. NO MORE
PROBLEMS  
and everything works fine.

Regards,
Artur.

On 2008-02-08, at 00:49, Todd Krause wrote:

>
> Hi,
>
> I'm new to this list and to lam-mpi, so forgive me if
I'm asking a  
> common
> question.  I tried to find the answer on the FAQs and
archives, but  
> either
> didn't find it, or didn't know it when I saw it.
>
> I've installed lam-mpi on my MacBook Pro, OS 10.5,
using Fink.  I'm  
> trying
> to use lam-mpi to run the downloadable cosmology code
Gadget2.  When  
> I run
> the program using mpiexec, I get the following output:
>
> 05:00 PM btpro:Gadget2> lamboot -v
>
> LAM 7.1.3/MPI 2 C++/ROMIO - Indiana University
>
> n-1<85442> ssi:boot:base:linear: booting n0
(localhost)
> n-1<85442> ssi:boot:base:linear: finished
> 05:01 PM btpro:Gadget2>
lamnodesn0	localhost:1:origin,this_node
> 05:01 PM btpro:Gadget2> lam-mpiexec -np 1 ./Gadget2
>
../bobtodd_tests/thick_disk_test01/thick_disk01_paramfiles/

> thick_disk01.param
>
------------------------------------------------------------
--------------
> Failed to find the following executable:
>
> Host:       Macintosh.local
> Executable: lam_appschema_ysCT98
>
> Cannot continue.
>
------------------------------------------------------------
--------------
> 05:01 PM btpro:Gadget2>
>
> I have not yet been able to figure out what this
problem is.  Any help
> would be greatly appreciated.  (I recently figured out
that, yes, I  
> should
> be using OpenMPI, bundled with 10.5... .  I'll cross
that bridge  
> once I
> get this ironed out. 
>
> Thanks very much,
> Todd
>
> PS -- output of laminfo follows.
>
> 02:46 PM btpro:Gadget2> laminfo
>              LAM/MPI: 7.1.3
>               Prefix: /sw
>         Architecture: i386-apple-darwin9.1.0
>        Configured by: root
>        Configured on: Thu Feb  7 14:48:44 CST 2008
>       Configure host: Macintosh.local
>       Memory manager: darwin7malloc
>           C bindings: yes
>         C++ bindings: yes
>     Fortran bindings: yes
>           C compiler: gcc
>         C++ compiler: g++
>     Fortran compiler: /sw/bin/gfortran
>      Fortran symbols: underscore
>          C profiling: yes
>        C++ profiling: yes
>    Fortran profiling: yes
>       C++ exceptions: no
>       Thread support: yes
>        ROMIO support: yes
>         IMPI support: no
>        Debug support: no
>         Purify clean: no
>             SSI boot: globus (API v1.1, Module v0.6)
>             SSI boot: rsh (API v1.1, Module v1.1)
>             SSI boot: slurm (API v1.1, Module v1.0)
>             SSI coll: lam_basic (API v1.1, Module
v7.1)
>             SSI coll: shmem (API v1.1, Module v1.0)
>             SSI coll: smp (API v1.1, Module v1.2)
>              SSI rpi: crtcp (API v1.1, Module v1.1)
>              SSI rpi: lamd (API v1.0, Module v7.1)
>              SSI rpi: sysv (API v1.0, Module v7.1)
>              SSI rpi: tcp (API v1.0, Module v7.1)
>              SSI rpi: usysv (API v1.0, Module v7.1)
>               SSI cr: self (API v1.0, Module v1.0)
>
> _______________________________________________
> This list is archived at http://www.l
am-mpi.org/MailArchives/lam/

_______________________________________________
This list is archived at http://www.l
am-mpi.org/MailArchives/lam/

It sounds like lam's mpiexec script is misbehaving somehow
-- you  
might try mpirun, instead (in LAM, mpiexec is a perl script
wrapper  
around mpirun).

Or you might try switching to Open MPI.  


On Feb 7, 2008, at 8:05 PM, Artur Tyliszczak wrote:

> Hi Todd,
>
> Better switch to OpenMPI as soon as possible.
> I spent couple of days installing and then trying to
run LAM 7.1.2 and
> 7.1.4 on the same computer and OS you have.
> I managed to install LAM, then I was able to compile my
parallel codes
> BUT there was no way to run them.
> Strange errors were reported and finally I gave up.
>
> Jeff Squyres suggested to go for OpenMPI. I did it. NO
MORE PROBLEMS
> and everything works fine.
>
> Regards,
> Artur.
>
> On 2008-02-08, at 00:49, Todd Krause wrote:
>
>>
>> Hi,
>>
>> I'm new to this list and to lam-mpi, so forgive me
if I'm asking a
>> common
>> question.  I tried to find the answer on the FAQs
and archives, but
>> either
>> didn't find it, or didn't know it when I saw it.
>>
>> I've installed lam-mpi on my MacBook Pro, OS 10.5,
using Fink.  I'm
>> trying
>> to use lam-mpi to run the downloadable cosmology
code Gadget2.  When
>> I run
>> the program using mpiexec, I get the following
output:
>>
>> 05:00 PM btpro:Gadget2> lamboot -v
>>
>> LAM 7.1.3/MPI 2 C++/ROMIO - Indiana University
>>
>> n-1<85442> ssi:boot:base:linear: booting n0
(localhost)
>> n-1<85442> ssi:boot:base:linear: finished
>> 05:01 PM btpro:Gadget2>
lamnodesn0	localhost:1:origin,this_node
>> 05:01 PM btpro:Gadget2> lam-mpiexec -np 1
./Gadget2
>>
../bobtodd_tests/thick_disk_test01/thick_disk01_paramfiles/
>> thick_disk01.param
>>
------------------------------------------------------------
--------------
>> Failed to find the following executable:
>>
>> Host:       Macintosh.local
>> Executable: lam_appschema_ysCT98
>>
>> Cannot continue.
>>
------------------------------------------------------------
--------------
>> 05:01 PM btpro:Gadget2>
>>
>> I have not yet been able to figure out what this
problem is.  Any  
>> help
>> would be greatly appreciated.  (I recently figured
out that, yes, I
>> should
>> be using OpenMPI, bundled with 10.5... .  I'll
cross that bridge
>> once I
>> get this ironed out. 
>>
>> Thanks very much,
>> Todd
>>
>> PS -- output of laminfo follows.
>>
>> 02:46 PM btpro:Gadget2> laminfo
>>             LAM/MPI: 7.1.3
>>              Prefix: /sw
>>        Architecture: i386-apple-darwin9.1.0
>>       Configured by: root
>>       Configured on: Thu Feb  7 14:48:44 CST 2008
>>      Configure host: Macintosh.local
>>      Memory manager: darwin7malloc
>>          C bindings: yes
>>        C++ bindings: yes
>>    Fortran bindings: yes
>>          C compiler: gcc
>>        C++ compiler: g++
>>    Fortran compiler: /sw/bin/gfortran
>>     Fortran symbols: underscore
>>         C profiling: yes
>>       C++ profiling: yes
>>   Fortran profiling: yes
>>      C++ exceptions: no
>>      Thread support: yes
>>       ROMIO support: yes
>>        IMPI support: no
>>       Debug support: no
>>        Purify clean: no
>>            SSI boot: globus (API v1.1, Module
v0.6)
>>            SSI boot: rsh (API v1.1, Module v1.1)
>>            SSI boot: slurm (API v1.1, Module v1.0)
>>            SSI coll: lam_basic (API v1.1, Module
v7.1)
>>            SSI coll: shmem (API v1.1, Module v1.0)
>>            SSI coll: smp (API v1.1, Module v1.2)
>>             SSI rpi: crtcp (API v1.1, Module v1.1)
>>             SSI rpi: lamd (API v1.0, Module v7.1)
>>             SSI rpi: sysv (API v1.0, Module v7.1)
>>             SSI rpi: tcp (API v1.0, Module v7.1)
>>             SSI rpi: usysv (API v1.0, Module v7.1)
>>              SSI cr: self (API v1.0, Module v1.0)
>>
>> _______________________________________________
>> This list is archived at http://www.l
am-mpi.org/MailArchives/lam/
>
> _______________________________________________
> This list is archived at http://www.l
am-mpi.org/MailArchives/lam/


-- 
Jeff Squyres
Cisco Systems

_______________________________________________
This list is archived at http://www.l
am-mpi.org/MailArchives/lam/

Thanks Artur and Jeff,

good suggestions -- both of them.  i tried mpirun, and it
worked (yes, you 
might think, well why didn't he try that before?  i *did*,
and it screwed 
up as well... but differently.  and then later i rebuilt
everything; and 
then again.  and *then* i forgot to try that... so a dumb
mistake, but i 
did try it back in the day).

okay, now that i got it to work, i can feel like i'm leaving
no loose ends 
as i try to get OpenMPI to work.  but that means i have to
get fftw2 to 
work by myself, since Fink configures it with LAM-MPI, not
OpenMPI.  but 
that gives me a whole litany of errors... and i'm using the
exact same 
installation procedure i used when i installed it
(correctly) before... 
but those were the days of OS 10.4.  guess that's the price
i pay for 
innovation.

oh, one other question: i created a .lamhosts file in my
home directory, 
which contains the one line

Macintosh.local cpu=2

is this file a) correctly written, b) in the right place,
given that i'm 
running lam in a totally different directory, and c)
irrelevant if i'm 
trying to run on only one computer (even though i'm hoping
to use both 
cpus)?

anyhow, thanks again to both of you.
best,
todd

On Thu, 7 Feb 2008, Jeff Squyres wrote:

> It sounds like lam's mpiexec script is misbehaving
somehow -- you
> might try mpirun, instead (in LAM, mpiexec is a perl
script wrapper
> around mpirun).
>
> Or you might try switching to Open MPI.  
>
>
> On Feb 7, 2008, at 8:05 PM, Artur Tyliszczak wrote:
>
>> Hi Todd,
>>
>> Better switch to OpenMPI as soon as possible.
>> I spent couple of days installing and then trying
to run LAM 7.1.2 and
>> 7.1.4 on the same computer and OS you have.
>> I managed to install LAM, then I was able to
compile my parallel codes
>> BUT there was no way to run them.
>> Strange errors were reported and finally I gave
up.
>>
>> Jeff Squyres suggested to go for OpenMPI. I did it.
NO MORE PROBLEMS
>> and everything works fine.
>>
>> Regards,
>> Artur.
>>
>> On 2008-02-08, at 00:49, Todd Krause wrote:
>>
>>>
>>> Hi,
>>>
>>> I'm new to this list and to lam-mpi, so forgive
me if I'm asking a
>>> common
>>> question.  I tried to find the answer on the
FAQs and archives, but
>>> either
>>> didn't find it, or didn't know it when I saw
it.
>>>
>>> I've installed lam-mpi on my MacBook Pro, OS
10.5, using Fink.  I'm
>>> trying
>>> to use lam-mpi to run the downloadable
cosmology code Gadget2.  When
>>> I run
>>> the program using mpiexec, I get the following
output:
>>>
>>> 05:00 PM btpro:Gadget2> lamboot -v
>>>
>>> LAM 7.1.3/MPI 2 C++/ROMIO - Indiana University
>>>
>>> n-1<85442> ssi:boot:base:linear: booting
n0 (localhost)
>>> n-1<85442> ssi:boot:base:linear:
finished
>>> 05:01 PM btpro:Gadget2>
lamnodesn0	localhost:1:origin,this_node
>>> 05:01 PM btpro:Gadget2> lam-mpiexec -np 1
./Gadget2
>>>
../bobtodd_tests/thick_disk_test01/thick_disk01_paramfiles/
>>> thick_disk01.param
>>>
------------------------------------------------------------
--------------
>>> Failed to find the following executable:
>>>
>>> Host:       Macintosh.local
>>> Executable: lam_appschema_ysCT98
>>>
>>> Cannot continue.
>>>
------------------------------------------------------------
--------------
>>> 05:01 PM btpro:Gadget2>
>>>
>>> I have not yet been able to figure out what
this problem is.  Any
>>> help
>>> would be greatly appreciated.  (I recently
figured out that, yes, I
>>> should
>>> be using OpenMPI, bundled with 10.5... .  I'll
cross that bridge
>>> once I
>>> get this ironed out. 
>>>
>>> Thanks very much,
>>> Todd
>>>
>>> PS -- output of laminfo follows.
>>>
>>> 02:46 PM btpro:Gadget2> laminfo
>>>             LAM/MPI: 7.1.3
>>>              Prefix: /sw
>>>        Architecture: i386-apple-darwin9.1.0
>>>       Configured by: root
>>>       Configured on: Thu Feb  7 14:48:44 CST
2008
>>>      Configure host: Macintosh.local
>>>      Memory manager: darwin7malloc
>>>          C bindings: yes
>>>        C++ bindings: yes
>>>    Fortran bindings: yes
>>>          C compiler: gcc
>>>        C++ compiler: g++
>>>    Fortran compiler: /sw/bin/gfortran
>>>     Fortran symbols: underscore
>>>         C profiling: yes
>>>       C++ profiling: yes
>>>   Fortran profiling: yes
>>>      C++ exceptions: no
>>>      Thread support: yes
>>>       ROMIO support: yes
>>>        IMPI support: no
>>>       Debug support: no
>>>        Purify clean: no
>>>            SSI boot: globus (API v1.1, Module
v0.6)
>>>            SSI boot: rsh (API v1.1, Module
v1.1)
>>>            SSI boot: slurm (API v1.1, Module
v1.0)
>>>            SSI coll: lam_basic (API v1.1,
Module v7.1)
>>>            SSI coll: shmem (API v1.1, Module
v1.0)
>>>            SSI coll: smp (API v1.1, Module
v1.2)
>>>             SSI rpi: crtcp (API v1.1, Module
v1.1)
>>>             SSI rpi: lamd (API v1.0, Module
v7.1)
>>>             SSI rpi: sysv (API v1.0, Module
v7.1)
>>>             SSI rpi: tcp (API v1.0, Module
v7.1)
>>>             SSI rpi: usysv (API v1.0, Module
v7.1)
>>>              SSI cr: self (API v1.0, Module
v1.0)
>>>
>>>
_______________________________________________
>>> This list is archived at http://www.l
am-mpi.org/MailArchives/lam/
>>
>> _______________________________________________
>> This list is archived at http://www.l
am-mpi.org/MailArchives/lam/
>
>
> --
> Jeff Squyres
> Cisco Systems
>
> _______________________________________________
> This list is archived at http://www.l
am-mpi.org/MailArchives/lam/
>
_______________________________________________
This list is archived at http://www.l
am-mpi.org/MailArchives/lam/

On Feb 7, 2008, at 11:25 PM, Todd Krause wrote:

> okay, now that i got it to work, i can feel like i'm
leaving no  
> loose ends
> as i try to get OpenMPI to work.  but that means i have
to get fftw2  
> to
> work by myself, since Fink configures it with LAM-MPI,
not OpenMPI.   
> but
> that gives me a whole litany of errors... and i'm using
the exact same
> installation procedure i used when i installed it
(correctly)  
> before...
> but those were the days of OS 10.4.  guess that's the
price i pay for
> innovation.

You need to recompile ffw2 with Open MPI.  LAM/MPI and Open
MPI are  
source compatible, but not binary compatible.

> oh, one other question: i created a .lamhosts file in
my home  
> directory,
> which contains the one line
>
> Macintosh.local cpu=2
>
> is this file a) correctly written,

Yes.

> b) in the right place, given that i'm
> running lam in a totally different directory,

In an rsh/ssh environment, the hostfile that you give to
lamboot can  
be anywhere; it's almost irrelevant what the name is. 
$HOME/.lamhosts  
has no special meaning (i.e., it's not a default location
that LAM  
looks for the hostfile).

> and c) irrelevant if i'm
> trying to run on only one computer (even though i'm
hoping to use both
> cpus)?

If you're only running on one host, it's irrelevant.  If you
do not  
specify a hostfile and LAM has nowhere else to get a
hostlist from  
(e.g., a resource manager), then LAM will assume that you
only want to  
run on the localhost.  You can then just "mpirun -np 2
...".

-- 
Jeff Squyres
Cisco Systems

_______________________________________________
This list is archived at http://www.l
am-mpi.org/MailArchives/lam/

great.  perfect.  thanks for all the help.

best,
todd

On Fri, 8 Feb 2008, Jeff Squyres wrote:

> On Feb 7, 2008, at 11:25 PM, Todd Krause wrote:
>
>> okay, now that i got it to work, i can feel like
i'm leaving no
>> loose ends
>> as i try to get OpenMPI to work.  but that means i
have to get fftw2
>> to
>> work by myself, since Fink configures it with
LAM-MPI, not OpenMPI.
>> but
>> that gives me a whole litany of errors... and i'm
using the exact same
>> installation procedure i used when i installed it
(correctly)
>> before...
>> but those were the days of OS 10.4.  guess that's
the price i pay for
>> innovation.
>
> You need to recompile ffw2 with Open MPI.  LAM/MPI and
Open MPI are
> source compatible, but not binary compatible.
>
>> oh, one other question: i created a .lamhosts file
in my home
>> directory,
>> which contains the one line
>>
>> Macintosh.local cpu=2
>>
>> is this file a) correctly written,
>
> Yes.
>
>> b) in the right place, given that i'm
>> running lam in a totally different directory,
>
> In an rsh/ssh environment, the hostfile that you give
to lamboot can
> be anywhere; it's almost irrelevant what the name is. 
$HOME/.lamhosts
> has no special meaning (i.e., it's not a default
location that LAM
> looks for the hostfile).
>
>> and c) irrelevant if i'm
>> trying to run on only one computer (even though i'm
hoping to use both
>> cpus)?
>
> If you're only running on one host, it's irrelevant. 
If you do not
> specify a hostfile and LAM has nowhere else to get a
hostlist from
> (e.g., a resource manager), then LAM will assume that
you only want to
> run on the localhost.  You can then just "mpirun
-np 2 ...".
>
> --
> Jeff Squyres
> Cisco Systems
>
> _______________________________________________
> This list is archived at http://www.l
am-mpi.org/MailArchives/lam/
>
_______________________________________________
This list is archived at http://www.l
am-mpi.org/MailArchives/lam/