OK, I'll take you up on the offer. I have 4 Power Mac G5's on a
private
network connected through a GigE switch. Even for large problems
the communications are slugish. This same code has shown to scale
to upwards of 128 processors on IBM SP's. So here is the output
to ompi_info --param btl tcp. Any help tunning this, or other
parameters
would be greatly appreciated. Are there any particular plugins for TCP?
-----------------------------------------------------------
MCA btl: parameter "btl_base_debug" (current value: "0")
If btl_base_debug is 1 standard debug is
output, if > 1 verbose debug is output
MCA btl: parameter "btl" (current value: <none>)
Default selection set of components for
the btl framework (<none> means "use all components that can be found")
MCA btl: parameter "btl_base_verbose" (current
value: "0")
Verbosity level for the btl framework (0 =
no verbosity)
MCA btl: parameter "btl_tcp_if_include" (current
value: "en1")
MCA btl: parameter "btl_tcp_if_exclude" (current
value: "lo")
MCA btl: parameter "btl_tcp_free_list_num" (current
value: "8")
MCA btl: parameter "btl_tcp_free_list_max" (current
value: "-1")
MCA btl: parameter "btl_tcp_free_list_inc" (current
value: "32")
MCA btl: parameter "btl_tcp_sndbuf" (current value:
"131072")
MCA btl: parameter "btl_tcp_rcvbuf" (current value:
"131072")
MCA btl: parameter "btl_tcp_exclusivity" (current
value: "0")
MCA btl: parameter "btl_tcp_eager_limit" (current
value: "65536")
MCA btl: parameter "btl_tcp_min_send_size" (current
value: "65536")
MCA btl: parameter "btl_tcp_max_send_size" (current
value: "131072")
MCA btl: parameter "btl_tcp_min_rdma_size" (current
value: "131072")
MCA btl: parameter "btl_tcp_max_rdma_size" (current
value: "2147483647")
MCA btl: parameter "btl_tcp_flags" (current value:
"10")
MCA btl: parameter "btl_tcp_priority" (current
value: "0")
_________________
Enrique Curchitser
ec...@columbia.edu
On Nov 19, 2005, at 1:04 PM, Jeff Squyres wrote:
Enrique --
I also recently added a section to the FAQ about TCP optimization.
It's a bit sparse at the moment, but please feel free to pester us
with
questions -- whenever we see good questions, we add them to the FAQ!
http://www.open-mpi.org/faq/?category=tuning
http://www.open-mpi.org/faq/?category=tcp
Be sure to look at the TCP BTL parameters that are available:
ompi_info --param btl tcp
Many of them are still lacking descriptions, so, again, feel free to
pester us with questions. :-)
On Nov 16, 2005, at 5:23 PM, Enrique Curchitser wrote:
Thank you. It did the trick! Now on to optimizing the
communications
over gig-ethernet.
_________________
Enrique Curchitser
ec...@columbia.edu
On Nov 16, 2005, at 4:29 PM, Troy Telford wrote:
On Wed, 16 Nov 2005 14:16:20 -0700, Enrique Curchitser
<ec...@columbia.edu> wrote:
Hi,
I put together a small cluster (4 computers) which has one head
node
that sees the world
and 3 that are on a private network. If I want to use the head
node
(which has 2 NICs)
as part of the ring, how do I tell it to go over the NIC that is
part
of the private network?
In mpich, there is a modification to the hostfile, but I did not
see
any such thing with OpenMPI
Here's an entry from an earlier post to this list (in the archives)
that I
believe applies:
http://www.open-mpi.org/community/lists/users/2005/11/0337.php
_______________________________________________
users mailing list
us...@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/users
_______________________________________________
users mailing list
us...@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/users
--
{+} Jeff Squyres
{+} The Open MPI Project
{+} http://www.open-mpi.org/
_______________________________________________
users mailing list
us...@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/users
