On top of that, and iirc, there used to be a bug in Open MPI that raised
some false warnings about how much memory can be registered.
OpenMPI 1.6.5 is antic now, so once you have updated your system
settings (see the faq, and ensure ulimit -l is unlimited on all your
nodes) and if still does
On May 18, 2016, at 6:16 PM, Ryan Novosielski wrote:
>
> I’m pretty sure this is no longer relevant (having read Roland’s messages
> about it from a couple of years ago now). Can you please confirm that for me,
> and then let me know if there is any way that I can silence
Hi there,
I’m getting the following message:
---
WARNING: It appears that your OpenFabrics subsystem is configured to only
allow registering part of your physical memory. This can cause MPI jobs to
run with erratic performance, hang, and/or crash.
This may be caused by your OpenFabrics vendor
Sorry, the figures do not display. They are attached to this message.
On Wed, May 18, 2016 at 3:24 PM, Xiaolong Cui wrote:
> Hi Nathan,
>
> I got one more question. I am measuring the number of messages that can be
> eagerly sent with a given SRQ. Again, as illustrated
Hi Nathan,
I got one more question. I am measuring the number of messages that can be
eagerly sent with a given SRQ. Again, as illustrated below, my program has
two ranks, rank 0 sends a variable number (*n*) of messages to rank 1 who
is not ready to receive.
[image: Inline image 1]
I measured
Folks --
In helping someone get up to speed with MTT yesterday, I updated an MTT sample
.ini file and accidentally committed the ompiteam-mtt Github password to the
public repository. :-(
I have therefore just changed the password for the ompiteam-mtt account.
Please contact me off-list for