[OMPI users] Exclude/Include HCA with OpenIB BTL ?

2007-07-05 Thread Don Kerr
Does the OpenIB BTL have the notion of include and exclude of HCA's as the TCP BTL does for NICs? E.G. "--mca btl_tcp_if_include eth1,eth2 ..." I think not but I was not sure if this was accomplished some other way so wanted to ask the group. TIA -DON

Re: [OMPI users] values of mca parameters whilst running program

2007-08-03 Thread Don Kerr
Glenn, If the error message is about "privileged" memory, i.e. locked or pinned memory, on Solaris you can increase the amount of available privileged memory by editing the /etc/project file on the nodes. Amount available (example of typical value is 900MB): % prctl -n project.max-device-locked-

Re: [OMPI users] memory leaks on solaris

2007-08-06 Thread Don Kerr
Glenn, With CT7 there is a utility which can be used to clean up left over cruft from stale MPI processes. % man -M /opt/SUNWhpc/man -s 1 orte-clean Achtung: This will remove current running jobs as well. Use of "-v" for verbose recommended. I would be curious if this helps. -DON p.s. or

Re: [OMPI users] memory leaks on solaris

2007-08-07 Thread Don Kerr
r stale daemons, then this will hopefully help (it isn't perfect, but it helps). However, orte-clean will not do anything about releasing memory that has been "leaked" by Open MPI. We don't have any tools for doing that, I'm afraid. On 8/6/07 8:08 AM, "Don Kerr&q

Re: [OMPI users] memory leaks on solaris

2007-08-07 Thread Don Kerr
s (which should kill their local apps) and will cleanup their associated temporary file systems. If you are having problems with zombied processes or stale daemons, then this will hopefully help (it isn't perfect, but it helps). However, orte-clean will not do anything about releasing memory that h

Re: [OMPI users] memory leaks on solaris

2007-08-10 Thread Don Kerr
n here. Orte-clean will kill any outstanding Open MPI daemons (which should kill their local apps) and will cleanup their associated temporary file systems. If you are having problems with zombied processes or stale daemons, then this will hopefully help (it isn't perfect, but

Re: [OMPI users] OMPI & uDAPL

2007-10-22 Thread Don Kerr
Couple of things. With linux I believe you need the interface instance in the 7th field of the /etc/dat.conf file. example: InfiniHost0 u1.1 nonthreadsafe default /usr/lib64/libdapl.so ri.1.1 " " " " should be InfiniHost0 u1.1 nonthreadsafe default /usr/lib64/libdapl.so ri.1.1 "ib0 0 " " "

Re: [OMPI users] mpirun udapl problem

2007-11-02 Thread Don Kerr
Also, you will want to check that you have valid entries in /etc/dat.conf. You might see the message you report if this is empty. Jeff Squyres wrote: Note that you still definitely need "self" in the BTL list. "self" is the loopback device/btl (i.e., one MPI process sending to itself; no

[OMPI users] Open MPI Linux Expectations

2008-05-22 Thread Don Kerr
Can anyone set my expectations with their real world experiences regarding building Open MPI on one release of Linux and running on another. If I were to... Build OMPI on Redhat 4, will it run on later releases of Redhat, e.g. Redhat 5? Build OMPI on Suse 9, will it run on later releases of Su

Re: [OMPI users] OpenMPI without IPoIB

2009-11-25 Thread Don Kerr
On 11/25/09 02:04, Vivek Satpute wrote: Hi, I tried to execute OpenMPI examples from OFED-1.4 stack and found that all those examples uses IPoIB. How are you determining that it is using IPoIB? Is there any way to execute MPI applications on only Infiniband stack i.e. without using IPoIB

Re: [OMPI users] does OpenMPI started deprecating udapl support

2010-01-07 Thread Don Kerr
I will take a look today. -DON On 01/07/10 09:34, Jeff Squyres wrote: This is a question for Sun to answer; they've been the primary/only drivers behind Open MPI's udapl support for quite a while. On Dec 30, 2009, at 2:22 AM, Nilesh Awate wrote: Hi all, I am new to Open MPI mailing li

Re: [OMPI users] does OpenMPI started deprecating udapl support

2010-01-07 Thread Don Kerr
Sorry I responded the wrong thread. Ignore my response. On 01/07/10 09:48, Don Kerr wrote: I will take a look today. -DON On 01/07/10 09:34, Jeff Squyres wrote: This is a question for Sun to answer; they've been the primary/only drivers behind Open MPI's udapl support for qui

Re: [OMPI users] btl udapl leaves string uninitialised

2010-01-07 Thread Don Kerr
Hi Dennis, I will look at this today. And apply the change as appropriate. -DON On 01/07/10 06:41, Dennis Schridde wrote: Hello! In btl_udapl_proc.c, if peer_proc->proc_addr_count <= 0, btl_addr_string will not be initialised. Attached is a patch that initialised btl_addr_string at the firs

Re: [OMPI users] does OpenMPI started deprecating udapl support

2010-01-07 Thread Don Kerr
Hi, I can only speak to Sun's use of uDAPL as it relates to MPI. From what I can tell Sun is the only group interested in the uDAPL BTL in Open MPI at this point. In contrast, Open MPI has many active members interested in the development of the OpenIB BTL which uses the user verbs protocol.

Re: [OMPI users] btl udapl leaves string uninitialised

2010-01-07 Thread Don Kerr
Yes I understand what you are doing but there is still a possible error case I was trying to consider and your initial placement of the call outside of the af==AF_INET check lead me to assume you were using something other than IPv4 which is why I was asking if you had an example. You don't a

Re: [OMPI users] UDAPL 2.0 support

2010-07-06 Thread Don Kerr
And Solaris has only implemented uDAPL 1.2. -DON On 07/06/10 08:00, Jeff Squyres wrote: We don't recommend using the udapl support in Linux; it is much better to use the native "openib" BTL that uses the verbs interface. We do not do any udapl testing on Linux, as far as I know -- the udapl

Re: [OMPI users] error in (Open MPI) 1.3.3r21324-ct8.2-b09b-r31

2010-07-15 Thread Don Kerr
There is a slightly newer version available, 8.2.1c at http://www.oracle.com/goto/ompt You should be able to install side by side without interfering with a previously installed version. If that does not alleviate the issue additional information as Scott asked would be useful. The full mpir

Re: [OMPI users] Heterogeneous OpenFabrics hardware

2009-01-26 Thread Don Kerr
Jeff, Did IWG say anything about there being a chip set issue?Example what if a vender, say Sun, wraps Mellanox chips and on its own HCAs, would Mellanox HCA and Sun HCA work together? -DON On 01/26/09 14:19, Jeff Squyres wrote: The Interop Working Group (IWG) of the OpenFabrics Allianc

Re: [OMPI users] 50% performance reduction due to OpenMPI v 1.3.2 forcing all MPI traffic over Ethernet instead of using Infiniband

2009-06-22 Thread Don Kerr
On 06/22/09 22:36, Jim Kress ORG wrote: Is there an environment variable (or variables) I can set to do the equivalent? OMPI_MCA_mpi_show_mca_params see: http://www.open-mpi.org/faq/?category=tuning#setting-mca-params Jim On Mon, 2009-06-22 at 19:40 -0600, Ralph Castain wrote: Sounds

Re: [OMPI users] Problem running MPI on a dual-core pentium D

2007-06-11 Thread Don Kerr
Additionally, Solaris comes with the IB drivers and since the libs are there OMPI thinks that it is available. You can suppress this message with --mca btl_base_warn_component_unused 0 or specifically call out the btls you wish to use, example --mca btl self,sm,tcp Brock Palen wrote: It

Re: [OMPI users] Problem running MPI on a dual-core pentium D

2007-06-11 Thread Don Kerr
Victor, You are right Prism will not work with Open MPI which Sun's ClusterTools 7 is based on. But Prism was not available for CT 6 either. Totalview and Allinea's ddt I believe have both been tested to work with Open MPI. -DON victor marian wrote: I can't turn it off right now to look

Re: [OMPI users] Problem running MPI on a dual-core pentium D

2007-06-11 Thread Don Kerr
the pentium is 4 time faster, but not 16 times. I wonder how a Sparc IV would perform. Victor --- Don Kerr wrote: Additionally, Solaris comes with the IB drivers and since the libs are there OMPI thinks that it is available. You can suppress this