--hpx:print-bind physical core numbers are wrong

[eschnett]

--hpx:print-bind outputs the following for me:

   0: PU L#0(P#0), Core L#0(P#0), Socket L#0(P#1), Node L#0(P#0)
   1: PU L#2(P#2), Core L#1(P#1), Socket L#0(P#1), Node L#0(P#0)
   2: PU L#4(P#4), Core L#2(P#2), Socket L#0(P#1), Node L#0(P#0)
   3: PU L#6(P#6), Core L#3(P#8), Socket L#0(P#1), Node L#0(P#0)
   0: PU L#0(P#0), Core L#0(P#0), Socket L#0(P#1), Node L#0(P#0)
   1: PU L#2(P#2), Core L#1(P#1), Socket L#0(P#1), Node L#0(P#0)
   2: PU L#4(P#4), Core L#2(P#2), Socket L#0(P#1), Node L#0(P#0)
   3: PU L#6(P#6), Core L#3(P#8), Socket L#0(P#1), Node L#0(P#0)

The physical core numbers look wrong; the sequence 0, 1, 2, 8 does not make sense. It seems the conversion from logical to physical core number (or from physical PU number to physical core number) is broken.

I started my application via

mpirun -x HPX_HAVE_PARCELPORT_TCPIP=0 -x MKL_NUM_THREADS=1 -np 2 ./bin/block_matrix --hpx:1:pu-offset=8 --hpx:threads=4 --hpx:pu-step=2 --hpx:print-bind

[eschnett]

I was confused. This system has 12 cores (not 8), and the reported numbers makes sense.