Josh Karns

`lstopo` output: ```bash Machine (63GB total) + Package L#0 Die L#0 L3 L#0 (8192KB) L2 L#0 (512KB) + L1d L#0 (32KB) + L1i L#0 (64KB) + Core L#0 PU L#0...

I compiled 4.1.4rc and still get the same output. It may be worth mentioning that my numa setup is sort of messed up since I only have two sticks of...

I noticed that if I run with 16 or fewer cores, it appears to work: `mpirun -x HWLOC_DEBUG_VERBOSE=1 -np 16 --report-bindings ./a.out 2> out` yields: ``` [server:369121] MCW rank 12...

Is there a more bleeding edge build of MPI I should try? Perhaps one that actually uses this so called PRRTE may avoid the issue.

Ive received those additional sticks of ram. If for some reason that resolves the issue we’ll have a good idea of what may be causing the bug? ill try v5.0.0rc...

Oddly enough, v5.0.0rc appears to work properly with two sticks (ran with 32 processes here): ``` [server:1555791] MCW rank 0 bound to package[0][core:0-7] [server:1555791] MCW rank 2 bound to package[0][core:0-7]...

So oddly enough now that each of my numa nodes actually has RAM, the split function appears to work as expected when I use mpi 4.1.4rc: `mpirun -x HWLOC_DEBUG_VERBOSE=0 -np...

Any thoughts on this? The bug still does exist, although the system configuration required to cause it is somewhat uncommon.

I can confirm that the work around used by @mjasyal-z works . It appears that the login for hangouts occurs in the station application where as the login for gmail,...

I like the idea of a FloatDecode trait, but it would effectively deprecate the current integer_decode method (which may be perfectly fine). I will work on an impl of the...