[mvapich-discuss] IME Support regression?

Subramoni, Hari subramoni.1 at osu.edu
Thu Sep 17 15:51:30 EDT 2020


Hi, Sergio.

Sorry to hear that you’re facing this issue. I have access to the bb5 cluster @EPFL. Could you please let us know how to reproduce the issue and we can look at it and get back to you.

Best,
Hari.

From: mvapich-discuss-bounces at cse.ohio-state.edu <mvapich-discuss-bounces at mailman.cse.ohio-state.edu> On Behalf Of Sergio Rivas-Gomez
Sent: Thursday, September 17, 2020 3:40 PM
To: mvapich-discuss at cse.ohio-state.edu <mvapich-discuss at mailman.cse.ohio-state.edu>
Subject: [mvapich-discuss] IME Support regression?


Good afternoon,

I work for the Blue Brain Project at EPFL, and I would like to report that I might have detected a possible regression in the IME support inside MVAPICH v2.3.4. I can confirm that both v2.3.2, which fixed an IME issue reported by our colleague Judit Planas, and v2.3.3 work perfectly fine as expected.

However, with v2.3.4, I am experiencing the following error (sharing only for rank #0):


    [0000] 2020-09-17 20:02:28 [error] (_im_topology_init:00162) Cannot set flags before loading topology [#topo]

    [0000] 2020-09-17 20:02:28 [panic] (im_get_obj_by_busid:00264) file: common/im-util/im_topo.c at line 264: 'im_assert(cpuset)'

Do you have any advice on what could be the issue?

Thank you very much in advance.

Kind Regards,
Sergio.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cse.ohio-state.edu/pipermail/mvapich-discuss/attachments/20200917/2b496fc3/attachment.html>


More information about the mvapich-discuss mailing list