Hi, I assume with the removal of `hwloc-assembler` and `hwloc_topology_set_custom` in 2.0+ that multi-node topology support has been relegated to netloc, but I figured I would run my use-case by you all just in case.
Essentially I want to do what `hwloc-assembler` did (i.e., combine multiple single-node topologies into one big one) with the 2.0+ API. No need for any complex network connections/topologies, a simple flat `system:1 machine:X core:Y PU:Z` topology will do. With that topology, I'd like to do two things: 1) print out a human readable summary equivalent to what you get with `lstopo-no-graphics` and 2) visualize the topology with `lstopo`. Is multi-node topology like this feasible with hwloc 2.0+? If so, are there any examples or documentation on how to achieve that? Thanks, Stephen
_______________________________________________ hwloc-users mailing list hwloc-users@lists.open-mpi.org https://lists.open-mpi.org/mailman/listinfo/hwloc-users