General

Profile

Jim Phillips

  • Email: jim@ks.uiuc.edu
  • Registered on: 01/23/2014
  • Last connection: 09/14/2017

Issues

Projects

Activity

09/18/2017

01:29 PM Charm++ Feature #1677: improved topology-aware partitioner
Yes, the code in Charm++ that handles the +partitions (+replicas) setup.

09/15/2017

04:33 PM Charm++ Bug #1676: Replicas slower than separate jobs on GNI on Blue Waters
Correct, as far as I know this is a GNI issue.
I've only tested on Blue Waters. It may or may not affect Titan, Eos...
09:56 AM Charm++ Bug #1676: Replicas slower than separate jobs on GNI on Blue Waters
verbs layer does not appear to be affected.

09/14/2017

11:28 AM Charm++ Feature #1677 (New): improved topology-aware partitioner
The current topology-aware recursive-bisection partitioner (described in Phillips et al., SC14) was based on an algor...
10:45 AM Charm++ Bug #1676: Replicas slower than separate jobs on GNI on Blue Waters
The bug does not affect the MPI layer on Blue Waters.
Still need to test verbs.
10:45 AM Charm++ 1.00 hour (Bug #1676 (New): Replicas slower than separate jobs on GNI on Blue Waters)
10:33 PM Charm++ Bug #1544: CMK_TIMER_USE_PPC64 inaccurate with variable clock speeds
What I was seeing in projections sure looked like it was just counting clock cycles.
10:31 PM Charm++ Bug #1674: Add 'ofi' target to autobuild
Bridges doesn't use 2-factor authentication.
09:59 PM Charm++ Bug #1676: Replicas slower than separate jobs on GNI on Blue Waters
I see the exact same performance for v6.7.0-574-g7d61794-namd-charm-6.8.0-build-2017-Jan-23-80737 and v6.7.0-0-g46f86...
09:33 PM Charm++ Bug #1676: Replicas slower than separate jobs on GNI on Blue Waters
6.8.0 from Sept 5 (v6.8.0-0-ga36028e-namd-charm-6.8.0-build-2017-Sep-05-28093).
No observed performance difference o...

Also available in: Atom